Abstract
Our life stories evolve in specific and contextualized places. Although our households may be our primarily shaping environment, our neighborhoods expose us to the immediate “real world” outside home. Indeed, the places where we are currently experiencing and have experienced life play fundamental roles in the development of our identities, beliefs, fears, perceptions of the world, and even in the (mis)understanding of our prospects of social mobility. Despite the analytic benefits that knowledge of the places and contexts where we experience life brings to better understanding our life stories and health outcomes, most qualitative, health, and mixed methods researchers continue to forego the analytic and elucidating power that geo-contextualizing our participants' narratives bring to our analyses, research findings, conclusions, and plans of action. From this view then, most qualitative, health, and mixed methods research findings and conclusions may have been ignoring the spatial contexts that most likely have shaped the experiences of research participants. The main reason for the underuse of these geo-contextualized stories is the requirement of specialized training in geographical information systems (GIS) and/or computer and statistical programming along with the absence of cost-free and user-friendly geo-visualization tools that may allow non-GIS experts to benefit from geo-contextualized outputs. To address this gap, we present GeoStoryTelling, an analytic framework and user-friendly, cost-free, multi-platform software that enables researchers to visualize their geo-contextualized qualitative narratives. The use of this peer-reviewed software (see https://doi.org/10.24433/CO.7385885.v3), available in Mac (https://cutt.ly/E724wAx) and Windows (https://cutt.ly/J724f0C) operating systems, does not require users to learn GIS nor computer/statistical programming to obtain state-of-the-art, and visually appealing and interactive maps. These resulting HTML visualizations are capable of integrating a variety of multi-media inputs that include text, image, videos, sound recordings/music, as well as hyperlinks to other websites—see https://cutt.ly/k7X9tfN. Since GeoStoryTelling relies on OpenStreetMap, its applications may be used worldwide, thus strengthening its potential reach to the qualitative, mixed methods, ethnographic, and health scientific communities, regardless of location around the world.
Keywords
Introduction
Our life stories evolve in specific and contextualized places. Although our homes may be our primarily shaping environment (via our parents or guardians), our homes are themselves situated in neighborhoods that expose us to the immediate “real world” or at least the version of the world where our homes were located (i.e., low-income neighborhoods look quite different than high income neighborhoods as we elaborate further below). This exposure to the places where we have experienced or are currently experiencing life, plays a fundamental role in the shaping of our beliefs, fears, perceptions of the world, and even our prospects for social mobility (Chetty & Hendren, 2018). In fact, this influence is so strong, that researchers have shown that most of us look like a weighted average of our childhood neighborhoods’ socioeconomic and demographic attributes (Chetty et al., 2014; Chetty & Hendren, 2018; González Canché et al., 2025). That is, if the norm in the neighborhood where we grew up was that its inhabitants held bachelors’ degrees, we were expected to have attained such a degree. On the other hand, if our childhood neighbors were prone to serve time in jail, this would have also been a likely possibility for ourselves or some members of our family and friends. In both instances (i.e., positive or negative neighborhood effects), the longer our exposure times to these neighborhoods were, the stronger these impacts have been estimated to be (Chetty et al., 2014, 2016; Chetty & Hendren, 2018). From this perspective then, a methodical effort to understanding the geo-contextualized information of the places where life happens, or stories evolve, may serve to better understand our overall perceptions of the world, and our outcomes. More to the point, geo-contextualization may help us reach a more nuanced understanding of our life stories, health and life outcomes, and decision-making processes.
Aside from serving as an introspective tool to better understand ourselves based on our life experiences, when applied to research, the incorporation of information that considers the attributes of the places where our research participants’ stories are happening, or have happened, may be instrumental in gaining a deeper and more nuanced understanding of their perceptions of the world and may help better inform the types of conclusions and implications we offer in our studies (Ajayakumar et al., 2019; Corbett & Keller, 2005; Jung & Elwood, 2010; Kwan & Ding, 2008; Matthews et al., 2005). More formally, the relevance of geo-contextualizing the stories of our research participants to better understand (a) their patterns and processes and (b) the overall influence of their environments in their behaviors and decisions (Mennis et al., 2013), explains the academic interest that health scientists have placed toward the overt incorporation of geospatial techniques and methodologies as part of the research process (Ajayakumar et al., 2019; Mennis et al., 2013). Indeed, “[h]ealth researchers incorporating spatial data have argued that we need to consider-the-local because of its importance in designing effective interventions” (Ajayakumar et al., 2019, p. 2). Here the effectiveness of the intervention is assumed to be strengthened based on the nuances gained when including geo-contextualized details. Note, however, that details in this context do not simply refer to place-based indicators, like in quantitative spatial modeling where it “suffices” to include poverty or income inequality in spatial regression models to “control” for spatial context. Instead, Ajayakumar et al. (2019) refer to the inclusion of context in the form or narratives of geostories, thus going well beyond merely adding layers with quantitative estimates of sociodemographic and socioeconomic place-based indicators. In their words, the interest is not only in indicating that something occurred, but more importantly on explaining why it did, or how it did (Ajayakumar et al., 2019).
The benefits of geocontextualization, however, are not limited to health research. In fact, most if not all qualitative and mixed methods studies, regardless of specific disciplines, may benefit from incorporating this geocontextualization as part of the research methodology implemented (Kwan & Ding, 2008). More to the point, we argue that geocontextualization does not need to be the primary analytic framework or methodology in a given study to strengthen our understandings. Even when used as an ancillary technique (i.e., playing a supportive role in our research endeavors), the inclusion of geocontextualized information may ultimately also lead to more nuanced conclusions and recommendations.
Despite the impact of the places where we experience life in shaping our perceptions of the world and outcomes, place-based narratives or geo-contextualized sets of beliefs and values that interpret our stories or accounts of events (Berengueres & Sandell, 2019), have typically remained untold and unmapped (Palis, 2022). That is, although it is evident that technologies like social media applications that can seemingly merge stories with location (i.e., geo-tagged pictures, videos, or posts) have been widely appreciated for over a decade (see Kwan & Ding, 2008; Matthews et al., 2005; Mooney et al., 2013), to date most qualitative, ethnographic, and mixed methods researchers continue to forego the analytic power that geocontextualizing ours and our participants’ narratives bring to our analyses, research findings, conclusions, and plans of action (Ajayakumar et al., 2019).
If Geo-Contextualization is so Important, why is it Not More Prevalent?
The main reason for the underuse of geo-contextualization in qualitative and mixed methods research in the social sciences, in relation to the natural and hard sciences (see Berengueres & Drozdzewski, 2024), is that, so far its application has required extensive specialized training in geographical information systems, geo-referencing and mapping, web-development, computer programming, and data science and visualization techniques. Moreover, as Ajayakumar et al. (2019) mentioned, it remains unclear and even daunting how to (a) map the stories of our participants to the location on earth where those events took place and were transmitted to us as our participants’ narratives (see Berengueres and Sandell (2019) for a longer discussion about data, stories, and narratives), (b) how to further include multi-media attributes that may take the form of texts, stories or narratives, images, audios, or videos, and (c) how to ultimately obtain fully interactive and multi-media HTML outputs that capture these now georeferenced stories.
Relatedly, another remarkable challenge is the current absence of free-to-use, user-friendly, and multi-platform software (i.e., available in Windows and Mac operating systems) capable of transforming these multi-media data into interactive HTML mapped outputs with absolutely no computer programming experience and/or geographical information systems (GIS) training. That is, currently available software like ArcGIS StoryMaps (https://storymaps.arcgis.com/) starts at a cost of $550 per year and still requires training in GIS. Mapbox’s Storytelling (https://www.mapbox.com/solutions/interactive-storytelling) allows users to “pay-as-they-go” but can get expensive quite fast (see https://www.mapbox.com/pricing), with starting prices as low as $16 that may reach $3200 per month. Not only these prices represent important investments, but also, when we consider that qualitative and mixed methods researchers tend to have access to less grant money resources than quantitative researchers, these prices may expand “digital and across disciplinary divides” (Drozdzewski & Berengueres, 2024, p. 1) therefore leading to lack of digitalization in disciplines with less access to financial and research grant resources (Drozdzewski & Berengueres, 2024).
More to the point, Mapbox’s Storytelling requires both training in GIS as well as statistical and computer programming or coding skills. A recent open-source software tool that integrates natural language processing, in addition to adding location pins to these written narratives is called WordMapper (Ajayakumar et al., 2019). 1 WordMapper, by itself, does not require programming skills nor specialized data formats (i.e., shapefiles storing points, lines, and polygons that are geo-located), and is free to use. However, although the use of WordMapper does not require specialized training, the data preparation may pose challenges. For example, in the case study presented by Ajayakumar et al. (2019), they required specialized hardware equipment for video and geo-tagging, as well as specialized software (Contour Storyteller https://contour.com/pages/storyteller) for data preparation. It is this use of specialized software that requires specialized training, which may be a barrier for the broad adoption of WordMapper among the qualitative, health, and mixed methods communities. Finally, in terms of its scope regarding operating systems (OS) and its multi-media handling capabilities, WordMapper is only available for the Windows OS and in its current version it does not accommodate multi-media formats such as image, video, audio, or external links.
Study Purpose and Contribution
As noted above, despite the relevance that geocontextualized information brings to qualitative and mixed methods research, the current and prevalent lack of access to multi-platform, free to use, and user-friendly software that requires no specialized hardware for data collection or preparation 2 remains a significant barrier for the widespread use of multi-media, geo-coded narratives, or stories (Drozdzewski & Berengueres, 2024) in social, mixed methods, and health sciences research. Accordingly, the purpose of this study is to fill these methodological and software-related gaps by introducing GeoStoryTelling, a completely cost-free, multi-operative system platform, and user-friendly geo-visualization methodology and Code Ocean (https://codeocean.com/) peer-reviewed software (https://doi.org/10.24433/CO.7385885.v3) that allows non-geographical information systems experts to benefit from the resulting fully interactive geovisualizations and multi-media HTML geonarratives.
Specifically, GeoStoryTelling, is both an analytic framework or methodology (i.e., a well-specified set of analytic steps to reach interactive outputs as we show in Figure 1) and a no-code and cost-free software that renders state-of-the-art, aesthetic and visually appealing interactive HTML maps that support and integrate multi-media inputs (i.e., text, image, videos, sound recording or music, and links to other websites) to provide further context to the geo- or place-based stories we are sharing. From this perspective, the overarching purpose of this study consists of providing a detailed step-by-step description of every component of the methodology and providing full and unrestricted access to the software tool available in Mac and Windows operating systems.
3
Flow Diagram for the Entire Methodological Process, Including Database Construction.
Another vitally important goal of this study consists of providing a detailed account of every step of the data building process where we demonstrate that no specialized software or hardware is needed to create our databases. Nonetheless, in addition to guiding researchers interested in using this methodology and software tool to build their own datasets, we are also providing a toy database (preloaded with the software but also available at https://cutt.ly/peFeyXEJ) so that researchers may interact first-hand with GeoStoryTelling and, while familiarizing themselves with this tool, they may also fully replicate the outputs discussed in this study.
The multi-media and place-contextualized narratives that can be shared via GeoStoryTelling, handle maps with a worldwide scope. As a map provider we use OpenStreetMap.org, which is open source, free to use, and it is available in 96 languages (see https://www.openstreetmap.org). In this respect, although the production of the HTML output requires an active internet connection to access these maps free-of-charge, GeoStoryTelling does not load our databases to any server, but all back-end processes are computed locally. In other words, we download maps from OpenStreetMap.org but we do not load our data to any server to conduct the operations, which naturally protects the privacy of our data inputs—having noted this, researchers have the possibility of posting their findings in servers like GitHub or RPubs, but they should only do so when there are
The remainder of this study is organized as follows. We next present a brief summary of the conceptual underpinnings of spaces and places, and their relevance for GeoStoryTelling. Subsequently, we present a summary of previous literature on geo-narratives and geo-stories, with an emphasis on depicting how GeoStoryTelling, as a methodology, contributes to this movement. In the next section we describe the flow diagram guiding the complete execution of GeoStoryTelling, from data repository and database creation to the execution of the back-end processes. Accordingly, our description of the flow diagram illustrates the data gathering and preparation processes and discusses the types of decisions researchers may need to make when preparing their databases. In this respect, we have paid particular attention to detailing the steps that researchers need to follow to create their own databases that can then be fed to the GeoStoryTelling’s user interface. Notably, the creation of these databases does not require any form of specialized software and can be compiled with free to use available tools, like Google Maps (https://www.google.com/maps), Bing Maps (https://www.bing.com/maps), or GPS Visualizer (https://www.gpsvisualizer.com), for example. After this discussion we present the GeoStoryTelling user interface (UI) and execute GeoStoryTelling with a pre-loaded toy dataset we are providing so that users may replicate all the outputs presented in this study. Finally, we describe the outputs and close with conclusions, limitations, recommendations, and future steps.
Conceptual Underpinnings and Prior/Current Applications
Where we grew up shaped or at least impacted our experiences. Take a moment and think about your childhood neighborhood. What do you see? How clean does it look? How does it smell? What sounds do you hear? Are there trees? Are there parks? Is it safe to walk around? What about public libraries? How far was the closest college? As discussed in our introductory section, the answers to these questions to a great extent shaped how you experienced life growing up in those neighborhoods and the types of memories you created, over and above the specific experiences and memories you had and created inside your home (González Canché, 2022; 2023; 2023, 2025). Even as an adult, these constant exposures to our neighborhoods continue to shape our experiences and behaviors. If we live in a dangerous or unsafe neighborhood, we would not feel comfortable walking late at night and will probably be extra careful with home security. On the other hand, some neighborhoods feel safe during any time of the day (González Canché, 2023).
Our quest to geo-contextualize our storytellings requires the blending of spaces and places, which although are eventually merged in our geo-stories may also be decomposed for further clarity, particularly during the data construction process. Spaces, and more specifically geographical spaces, are inherently physical and therefore objectively measurable, identifiable on, and delimitable around Earth’s surface (Cresswell, 2008; Tuan, 1977). Spaces are also abstract (Tuan, 1977) for they may be represented with grids or systems of mathematical coordinates (Agnew, 2005; Agnew & Livingstone, 2011) that are not necessarily earth bounded (i.e., we can map our brain activities or measure distances across galaxies following these same mathematical principles). However, when this grid or coordinates system is applied to earth’s surface, we refer to spaces as being geolocated or georeferenced for the geo root or prefix comes from the Greek meaning of earth or g
In GeoStoryTelling, georeferenced spaces are relevant for they will enable us to geolocate the narratives or stories we seek to contextualize. However, mathematical concepts (i.e., spaces) without context limit our abilities to understanding social experiences (González Canché, 2023). It is this goal or challenge of contextualizing abstract concepts, where
This simple equation that renders splaces, captures the mental process we followed when asked to remember our childhood neighborhoods. That is, as part of the exercise of retrieving our memories associated with those childhood places, we may also have remembered measurable attributes such as distances we may have traveled from home to school along with other physical attributes that may have affected our experiences associated with those places (i.e., temperature, elevation). These attributes may also go beyond physical ones and may include socially based indicators like poverty, crime, and educational attainment levels. From this perspective then,
This brief discussion aimed to elaborate on how places and spaces intermingle on a daily basis forming
GeoStoryTelling Through the Splace-Time Continuum
Our experiences are always evolving through the splace-time continuum. Accordingly, our GeoStories must not be constrained to specific times, instead, our analyses should freely consider our participants’ experiences in a time and splace that is meaningful to them and their experiences. For example, assume we want to understand the life path or trajectory followed by a scholar who despite having grown up in poverty achieved to become a professor in a worldclass university. In this case, the GeoStory may begin by depicting the context where and when this scholar grew up. Describing her childhood neighborhood and then moving to her experiences during college, situating this description in the time and area where she attended college and then moving to graduate school and to her academic employment trajectory, documenting the multiplicity of employments she has held since becoming a faculty member. In this example, the resulting visualization will locate these stories across multiple splaces (i.e., earth’s surface via latitude and longitude coordinates that formed meaningful memories) while covering several decades in time. In this case, it is our responsibility as GeoStoryTellers to provide clarity and enough splace and time context as part of our depiction of our participant’s trajectory. To achieve this clarity, we could start by adding an ID or name to each geo-narrative that clearly conveys the time when the event took place along with some meaningful yet succinct description that will be added as a legend to the resulting GeoStory—with this information the resulting GeoStory map will result in a combination of a set of unique colors that will be added as pins to the interactive map to ease our navigation and enhance our storytelling (see Figure 2, for example). In addition, the GeoStoryTelling software allows us to add video, audio, or hyperlinks to other websites that illustrate her academic trajectories, like the types of colleges she attended or the types of neighborhoods where she grew up. This information may be added in addition to the narratives we selected to share with each point in her trajectory map—we will illustrate all these mapping and visualizing possibilities below. Example of GeoStoryTelling, HTML Interactive Version Available at https://cutt.ly/k7X9tfN.
By a similar token, our maps may present the GeoStories of different people and events. For example, in the HTML map that may be replicated with the toy database and the sotware tool we are providing (see https://cutt.ly/k7X9tfN and Figure 2), we include six points that illustrate different and disconnected stories. This disconnection is strategic for it serves to showcase the versatility of GeoStoryTelling in capturing multiple types of narratives to be shared. For example, one of these stories is a behind the scenes of the film called “Kensington Blues” https://kensingtonblues.com/, which documents a drug epidemic in Philadelphia. Another story is a presentation that took place in late 2016 on spatial and network analyses applications at the University of Pennsylvania. Yet another story showcases an audio recording of a 2022 concert “Avalon Emerson-Live.” Additionally, three locations showcase examples of places where traditional and non-traditional cultural and literacy-oriented activities take place. These cases account for a home, a cultural center for (mostly) undocumented families, and a well-established museum. What is common across these stories is that we demonstrate the inclusion of hyperlinks to find more information about each case. For example, in the case of Kensington we provide a link to a Time Magazine report on the opioid crisis https://time.com/jeffrey-stockbridge-kensington-philadelphia-photographs/, in addition to the possibility of playing the documentary directly from the GeoStoryTelling interactive map.
As can be seen in Figure 2, the resulting HTML map (see https://cutt.ly/k7X9tfN) includes a legend key that enables us to easily locate the GeoStory of interest to display more information, including the multimedia inputs. That is, the legend key reveals more information about each point based on attributes of the name column. Users may decide to add categories instead of unique names, also under this column. For visualization purposes, Geostories located in close distances are grouped together based on a cluster algorithm based on a radius of 80 pixels, or roughly 1.3152E-5 miles. That is, the number “3” in Figure 2 indicates that three events were clustered together. Clicking in each cluster will automatically zoom in revealing the points located in such a cluster. If events took place in the same location, GeoStoryTelling will display a localized geo-network revealing each GeoStory. Users may refresh the browser to reset the HTML or hit the reset button located in the top left side of the map. Note that one of the stories shared in this map, deals with the opioid epidemic, which may be triggering, accordingly, viewer discretion is advised (as indicated in the legend).
Although we describe in more detail this output when discussing the implementation of GeoStoryTelling, in the following section we present a summary of how similar methodologies have been discussed and implemented in the existing literature. This discussion also allow us to highlight how GeoStoryTelling is contributing to this line of inquiry and methodology.
How Have GeoStories or GeoNarratives Been Implemented?
Two of the first methodological discussions of how to integrate GIS with narratives I was able to find, were published by Matthews et al. (2005) and Corbett and Keller (2005). In their study Matthews et al. (2005) sought to merge or integrate GIS with ethnographic data relying on a multi-site study of low-income families and their children. With this integration, referred to as geo-ethnography, the authors aimed to better situate the actions and experiences of their participants in time and space to enhance data analysis and interpretation. Like in GeoStoryTelling, Matthews et al. (2005) also elaborated on the relevance of further incorporating other data inputs such as photographs, audio, and video files. An important cautionary note raised by these authors, that we also share with our GeoStoryTelling approach, is that the resulting visualizations, although useful are not the end analytic product, but just a mean toward gaining a deeper understanding. In their words “[w]e still need ethnography to help us piece together how many of the contextual datasets that can be spatially joined are relevant to low-income families and low-income communities (Matthews et al., 2005, p. 70). Similar to the StoryMaps (https://storymaps.arcgis.com/) description in the introduction, Matthews et al. (2005) wrote an extension to be executed in the proprietary software ArcGIS (https://www.esri.com/), thus having potentially high cost and requiring specialized training in geographical information systems—GIS.
Like Matthews et al. (2005), Corbett and Keller (2005), also demonstrated their integration of GIS with qualitative data in a study with disadvantaged groups. With the goal of discussing notions of empowerment versus marginalization, Corbett and Keller referred to their integration as participatory GIS. The authors argue that although discussion of integrating GIS with social and community research has been traced since at least 1995, then current applications were referred to as participatory GIS (or PGIS) (Corbett & Keller, 2005). The term PGIS was adopted with the goal of combating the idea that GIS by itself was becoming “a positivistic and technocratic tool that supported the more powerful sectors of society, often at the expense of weaker groups” (Corbett & Keller, 2005, p. 92). From this perspective, GeoStoryTelling ascribes to this PGIS notion by providing a methodology and software tools that require absolutely no monetary cost nor specialized training. That is, like PGIS, GeoStoryTelling seeks to empower qualitative and mixed methods researchers—and the general public—, who despite not having had the opportunity to receive specialized training may benefit from these open-source data science and visualization technologies.
One of the most cited papers on this topic (Kwan & Ding, 2008) also aimed to extend GIS capabilities for the analysis and interpretation of narrative materials, including oral histories and biographies. In their discussion Kwan and Ding (2008) argued that the use of GIS methodologies may be used to “complement or triangulate [(i.e., verify results using multiple data sources)] the knowledge acquired through the qualitative component of the research” (Kwan & Ding, 2008, p. 444). Like in the cases of Carroll (2020) (with StoryMaps) and Matthews et al. (2005), Kwan and Ding (2008) also developed and analytic component in ArcGIS (https://www.esri.com/). Similarly, it is likely that the Wordmapper software developed by Ajayakumar et al. (2019) was inspired by Kwan and Ding’s description of the need to incorporate text processing capabilities into GIS. For not only Ajayakumar et al. (2019) cited Kwan and Ding (2008), but Ajayakumar et al. also included natural language processing capabilities in the form of world clouds in their use of geonarratives.
More recently, Mennis et al. (2013) introduced a prototype implementation of an interactive, cartographic visualization software application for exploring qualitative activity space data that was also based on geonarratives. They referred to this tool as Qualitative Activity space data Viewer (QAV). Although the description of this tool shows important features incorporated in GeoStoryTelling, like the integration of multi-media data, there was no information on how to access that QAV prototype. Similarly, an extensive web search generated no results on the existence of this software. From this perspective, the open-source software provided by Ajayakumar et al. (2019) is perhaps the closest cost-free and no-code option to GeoStoryTelling that we were able to locate. Unfortunately, as of April of 2023, we were not able to get this software to work.
Two alternatives to Wordmapper (see Ajayakumar et al., 2019) that have become popular options based on their support of multi-media inputs for the narratives are: Story Maps (Carroll, 2020) and Mapbox’s Storytelling (https://www.mapbox.com/solutions/interactive-storytelling). Despite their popularity, these software tools require specialized knowledge (i.e., GIS, coding, or both) and may become quite expensive.
The most recent development on this movement was offered by Li et al. (2023) and was featured in the 2023 Conference on Human Factors in Computing Systems. This tool, called GeoCamera, aligns with the goal of GeoStoryTelling to lessen the barriers in incorporating geographical information system data but instead of textual narratives, audio, and images, GeoCamera was developed to tell stories with geopraphic data videos. That is, GeoCamera seeks to produce compelling camera movements for storytelling. In short, as Li et al. (2023) stated, GeoCamera aims to lower the barrier of crafting diverse camera movements for geographic data videos and its user interface was designed to empower users to tell appealing geographic stories with tailored camera movements. GeoCamera renders visually appealing outputs, however, we were unable to get access to the user interface.
One of the reviewers of this paper pointed us toward two AI tools (see Anthropic and v0.dev) capable of generating standalone HTML maps like those generated by GeoStoryTelling with prices starting at $20 USD per month. The input data for these maps is a database like ours that needs to be uploaded to their servers, therefore potentially raising issues with data privacy. As briefly mentioned above, a notable, free to use alternative to generate geo-coded data that can be loaded to Google Earth is QualNotes (see Drozdzewski & Berengueres, 2024), which can easily be integrated with GeoStoryTelling. QualNotes is capable of geo-tagging qualitative data in the form of participant observations, mobile mapping, and audio and video interviews.
Aligned with the empowering notions depicted by Corbett and Keller (2005), Li et al. (2023), and Drozdzewski and Berengueres (2024), GeoStoryTelling offers access to user friendly, no-code, and no-cost software, that, although builds from state-of-the-art mapping and interactive visualization options, requires no training in GIS to link texts or narratives to their abstract spaces—in the form of longitude and latitude coordinates. Moreover, GeoStoryTelling was designed with the purpose of adding “hot-links” as described by Matthews et al. (2005). These “hot-links” are live hyperlinks (i.e., active text or graphic link on a Web page) that allow users interacting with GeoStoryTelling to explore other websites where more information about the narratives of interest may be gained. Additionally, these “hot-links” may also display audio, images or photographs, and videos directly in the interactive map rendered by GeoStoryTelling. All without any need to code or pay for specialized equipment or software.
In sum, this subsection described applications related to GeoStoryTelling. Notably, as depicted here, our proposed methodology and software tool stands as the only no-code, non-technical, multi-platform that is completely free to be used and to be distributed. Given the potential salience and impact of GeoStoryTelling, the following section details the set of specific steps that users may follow to build the databases that GeoStoryTelling will then transform into interactive HTML outputs. It is with this end, that we now proceed to discuss the methodological steps as described in the flow diagram operationalized in GeoStoryTelling.
Flow Diagram for the Entire GeoStoryTelling Process
Figure 1 represents the complete set of steps required to execute GeoStoryTelling—from data repository creation to software execution. With the purpose of providing as much transparency and clarity of the entire process, every step (or polygon) in the flow diagram presented in Figure 1 may be conceptualized as follows:
1. Collect stories or narrative and select text to be featured in the GeoStoryTelling interactive visualization.
2. Begin the data repository creation by placing those selected texts in a comma separated value file in Google Sheets or Microsoft Excel, for example. As we further explain below, the database to be uploaded to GeoStoryTelling only requires a maximum of seven columns. a) The column containing the ID or name of the story will be used to add a legend key to the interactive map. If repeated names are used all those stories will share the same color, but if different IDs are provided, each resulting pin in the map will have a unique color, thus easing the identification of GeoStories and the navigation in the HTML interactive map.
3. Situate those texts on Earth’s surface (i.e., geocode or georeference them) by adding latitude and longitude coordinates.
4. Although optional, we recommend adding the so called “hot-links” to other websites with further information about the texts selected as well as video, audio, or images links. Note that images of photographies can also be loaded from your local computer or may be taken from a live link from the web.
5. Load the database to the GeoStoryTelling user interface available in Mac and Windows operating systems that run all processes locally—hence no need to upload your data to any server.
6. Select your columns with a few clicks. When databases are loaded, the user interface automatically detect the column names, which can be selected with a few clicks.
Illustration of the Steps for Database Creation (Format is a Comma Separated Value [.csv] File).
aAs we explain in the content of the study, this link is called embedded for videos and audios, see Figure 4(a) and 4(b).
bIn the cases of images, they can be read from our local computers or from an image hyperlink.
cThe only required elements to create maps are the contents of longitude, latitude, and geostories. All other elements are optional and of these elements some may be semi-populated. For example, the text in brackets indicates that while two of the three links to webpages are populated, one is not. The HTML outputs are then flexible to handle missing data.
8. The resulting HTML map may be shared via email or hosted on a website, like GitHub or RPubs for free (see our fully reproducible example here https://cutt.ly/k7X9tfN). However, researchers need to be completely sure that no privacy or security concerns exist if they decide to upload these HTML files.
To provide as much clarity as possible, each of these steps is further detailed as a subsection in the following pages.
Stories or Narratives Collection
The process starts with the collection of stories or narratives. The data gathering of these stories may follow an ethnographic, qualitative, survey, or mixed methods approach. These stories may also take the form of interview transcripts, essays or reflections, open-ended responses in surveys (see the “Application Example to Survey Responses Involving Campus Safety” subsection below), videographies (i.e., capturing moving images on electronic media and even streaming media)—like in the case of Ajayakumar et al. (2019) who collected videos and then transcribed the voices in those videos for their geonarratives—, ethnographic observations (Matthews et al., 2005), social media posts—i.e., like volunteered geographic information (Mooney et al., 2013)—, and more generally, any type of form of text or qualitative evidence collected (see Drozdzewski & Berengueres, 2024 as well).
In this respect, although GeoStoryTelling is capable of displaying vast amounts of text in the resulting visualizations, we strongly recommend researchers to be selective with and purposeful in the selection of the narratives or stories they choose to display. From this perspective, we also highly recommend they practice interactive reviewability (Alexander et al., 2019; González Canché, 2022, 2023a, 2023b, 2023c), where they ask their research participants whether the preliminary text selection conducted by the researchers is appropriately capturing their participants’ experiences, or whether there should be better narratives or aspects of their participants’ episodic memories (Murray, 2018) that more accurately capture or reflect their viewpoints and/or experiences. Relatedly, please note that GeoStoryTelling is a tool that aims to convey ideas and experiences. Accordingly, GeoStoryTelling should be viewed as a contextualizing tool, rather than as the end goal of our studies. That is, as Matthews et al. (2005) stated, this tool should not be the only source of analysis; instead, more qualitative evidence is still needed to reach well balanced and nuanced analyses and useful conclusions, recommendations, and plans of action.
Database Creation
Once we have gathered and selected the texts or narratives that we want to feature in our GeoStoryTelling, we are then ready to start creating our database. This process is as straightforward as adding seven columns to a Microsoft Excel or Google Sheets object and save it as a comma separated value (or *.CSV) file or database. Note that this process will be the same if we have large scale surveys, as we discuss in the subsection called “Application Example to Survey Responses Involving Campus Safety.”
As we proceed with our discussion, we will elaborate more on each of these seven columns. Note that we do not have to follow the column names, nor the order, we are showing below and in Table 1. However, for the sake of clarity let us present the seven column names we are displaying in our database example. These names are: 1. Name or ID 2. Location or Address 3. Link to Webpage 4. Video, Image, or Audio Link. Once more, images may be loaded from our local computer or from the internet. In the case of videos and audios we need to obtain the embedded versions of the corresponding links, as we illustrate below. 5. Longitude 6. Latitude 7. GeoStory
In this list of columns, GeoStory (i.e., the text or narrative we just discussed above) is the last column of that database. The seventh or last position of this column is just selected for convenience. The text content of this column will typically result in wider columns than most of the other column contents, which may complicate readability if GeoStories are placed earlier in the database. Having said this, as we mentioned earlier in this subsection, the order of the columns is not important for the GeoStoryTelling software to render the HTML interactive outputs.
To illustrate the step-by-step process followed to build a database that can be used with GeoStoryTelling, we have added four panels to Table 1, each of which can be mapped back to our flow diagram shown in Figure 1.
Panel A in Table 1 shows the initial status after the addition of the seven columns, we described above, has been completed. Panel B represents the addition of the narratives or stories that we want to feature. For illustrative purposes, we are simply populating these cells with generic texts that read “1st story text example,” ‘2nd story text example,” finishing with a “Nth story text example.” That is, our database may accommodate as many stories as we need or want to feature. The addition of these cases is as easy as copying each of the texts selected from our transcripts and paste them in the column called “GeoStory” (once more note that column names do not need to adhere to the names I am using in our working example).
Also note that in Panel B in Table 1 we have also added identifiers to each of these stories. To illustrate the flexibility of GeoStoryTelling, we decided to feature the story of the same participant (called “A”) at different points in time in the first two rows. This is participant A, whose first selected story took place at time 1, and may be represented as IDA,t = 1. Moreover this same participant A also provided us with another story, which in this case took place at time 2 (IDA,t = 2). This representation of time follows global time stamps, as described by Artale et al. (2007) and Jensen et al. (1998)—see also González Canché (2022) and González Canché (2023). Global time stamps allow us to be flexible in how time is measured. That is, time may be measured in months for participant A in one instance, and it may also be measured in years for that same participant in her second story, based on her storytelling or episodic memories (Murray, 2018). Moreover, these time elements may not be linear in how they evolved, instead, the storytelling of our participants may jump back and forth in the space-time continuum, and it is our challenge and responsibility as researchers to capture these jumps—and to consult with our participants if our interpretations are accurate.
Non-linear temporal dimensions in qualitative, ethnographic, and mixed-methods research enable more flexibility in capturing nuanced understandings of the stories our participants are sharing (Neale, 2015, 2019). More specifically, and organically speaking, what our participants shared regarding “time X,” may have prompted them to share more information regarding “time X + n,” and this latter story may have prompted them to share events that occurred at “time X−n.” Here, X is a point in time situated around our participants’ experiences, and n represents time units before or after such a point in time took place. Once more, the n time units may be hours, days, years, or decades, for example, before or after a given point in “time X.” To minimize confusion and create common understandings, let us assume that a participant shared with us the following narrative: “I finished college in
Here, this participant offered information initially situated in 1997, but then there was a jump in time ten years into her future with respect to 1997 (time X + n), when more information about graduate education was offered. Also note that in this same example, our research participant followed her trend of thought by jumping back to her high school years, which, following this non-linear time rationale, may be expressed as “time X−n.”
This brief example serves to showcase two main features. We can accommodate multiple mini-stories of our participants and the units of time we may depict do not need to follow linearly evolving trends nor to even be measured in the same unit of time. Instead, we may capture more nuance in the storytelling of our participants by applying non-linear, global time stamps. In addition to enable the inclusion of meaningful identifying information, the content of the “Name or ID” column will be displayed as a legend to the HTML rendering and will be automatically mapped with specific colors to add more context and strengthen the navigation process. From this view, all events that share the same ID, will have the same color in the GeoStory.
As shown in Figure 1, the next step in our database building process consists of geo-referencing those stories or narratives.
GeoTagging or Situating Those Stories on Earth’s Surface
Note also that in this brief example, our participant shared with us information that may be situated on Earth’s surface. Specifically, she mentioned Mérida, Spain and Los Angeles, California. Although our participants’ stories will likely not be as short as these experiences shared in our example, recall that the selection of texts or narratives is based on the purpose of the study and decisions of the research team. For convenience, let us assume that IDA,t = 1 and IDA,t = 2 in Table 1, capture the graduation from college in 1997 and her starting of graduate education 10 years later (and in other country), respectively.
As described earlier, the only required elements that GeoStoryTelling need to produce the HTML outputs are latitude and longitude coordinates, which are the most fundamental pieces of information that enable us to locate units on Earth’s surface. Although coordinates may take two main forms, depending on whether we rely on a planar projection of Earth (measured in meters) or in a geodesic grid (measured in degrees) (see González Canché, 2023; Holdgraf & Wasser, 2022), for our GeoStoryTelling purposes, let us just note that we may retrieve these coordinates from services like Google Maps https://www.google.com/maps, Bing Maps https://www.bing.com/maps or GPS Visualizer https://www.gpsvisualizer.com. The latter is an online, free to use tool that conveniently allows us to retrieve batches of coordinates at the time (i.e., dozens, hundreds, or thousands of addresses or locations as we illustrate below). This process of retrieving latitude and longitude coordinates from addresses or locations is called geo-coding or geo-referencing—hence we will use these terms interchangeably henceforth.
Geo-coding or Geo-Referencing
To further elaborate on this geo-referencing process let us discuss Figure 3. This figure shows the result of retrieving the latitude and longitude coordinates using Bing and Google maps services. In each case, we may input an address in the search bar or simply right click on a desired point in the map. In Figure 3 we right clicked on the same location (i.e., 4341 Baltimore Avenue in Philadelphia, PA, USA). As you can see in both cases, we have two sets of numbers that, although show some variations (based on each service implementation of Earth’s projections), in the big scheme of things, match the representation of the location of interest. Geocoding or Georeferencing Processes.
The first number (starting with 39.94…) is the latitude that ranges from −90 to 90°. A value of 0 represents the Equator, which divides Earth into the north and south hemispheres. Positive values indicate that such a point is north from the Equator and negative values indicate that such a point is below the Equator (González Canché, 2023). The second number is the longitude coordinate. Longitudes range from −180 to 180°. A value of zero degrees longitude, is referred to as the Prime Meridian located in Greenwich. From the Prime Meridian to the left, up to
For our purposes, we just need to copy these coordinates and paste them in our database under the latitude and longitude columns shown in Table 1. Note that this georeferencing process works with specific or more general locations. That is, if we input addresses, the results will be quite precise. However, we can also just input or type city names or even country names and services like Google or Bing maps will still render the latitude and longitude coordinates capturing the centroid (center) of those cities or countries. In going back to our brief narrative, our participant mentioned two places: Mérida, Spain and Los Angeles, California. This information serves to demonstrate a batch geocoding approach also available free of charge when connected to Bing Maps via GPS Visualizer (https://www.gpsvisualizer.com/geocoder/). To use this service free of charge across the globe, we need to register at Bing (https://www.bingmapsportal.com/). This is a process that takes five minutes and results in a unique access key like the one we covered in red our rendering of Figure 3—note that we should never share our free Bing API key.
Batch geocoding is useful when we are dealing with multiple locations or large scale projects. That is, although in our brief example we only have two locations, the same rationale applies with hundreds or thousands of addresses or general inputs like cities and countries. As can be seen in Figure 3, we may select Bing as the geocoding engine and add our Bing API key. Then we simply need to copy and paste the addresses or places we need to georeference. This latter process takes two main forms. The first is a raw list with one address per line and a tabular form where the engine assumes that the first line contains column names and, accordingly, ignores the information contained in this first line for the batch geocoding procedures. For example, let us assume that we want to geocode a version of Table 1. If we copy and paste this table or the specific column containing the address or location to be geocoded and paste it into GPS visualizer, and select, “Type of data: tabular (columns & a header)” the process will render the same result as shown in Figure 3. But if we input a tabular data and select “Type of data: raw list, 1 address per line,” the batch geocoding process will try to geocode the line containing column names, which will result in missing cases for this first line—i.e., column names do not have coordinates, unless one of the column names is recognized as a place around Earth’s surface, like a column called “Rome,” for example.
The result of this batch geocoding process will be a comma separated output that can be copied into excel. This tabular output will contain the following elements: latitude, longitude, name, desc, color, source, precision. Name matches our input requests (i.e., Merida, Spain and Los Angeles, California, USA as shown in Figure 3) to ease the identification of the addresses. The column “desc” shows the actual name of that location as reported, in this case, by Bing. For example, our first input name was “Merida, Spain” and the “desc” column rendered “Mérida, Extremadura, Spain.” That is, Bing maps also rendered the region in Spain, where the city of Mérida is located.
As in the previous example, we need to take the latitude and longitude coordinates and add them to our database. By doing so, we have effectively been able to link our participants’ narratives to the spaces they are located on Earth’s surface. This result is shown in Panel C in Table 1, where we are adding both the location information and the corresponding latitude and longitude coordinates.
This description concludes the minimum elements required to produce interactive maps via GeoStoryTelling. However, although at this point the resulting maps will display the IDs, 4 the locations’ names, and the narratives, no other multi-media data have yet been incorporated in our GeoStoryTelling database. This brings us to our last subsection, which as described in Figure 1, is optional, but highly recommended to provide more detail and nuance to our GeoStories.
Additional Contextualizing Elements: Video, Audio, Images, and External Hyperlinks
Like the process of adding latitude and longitude, the addition of video, audio, or images that can be directly displayed along with our GeoStories, does not require specialized software or hardware. Instead we can retrieve these pieces of information from resources freely available in sites such as SoundCloud (https://soundcloud.com), YouTube (https://www.youtube.com/), vimeo (https://vimeo.com/), or Spotify (https://open.spotify.com/) to mention a few streaming services. In the case of images, we can feature any image hosted on online by simply retrieving its direct link from the internet—as we show below. Alternatively, we can also load any image stored on our computer to be displayed as part of our GeoStoryTelling.
Note that, although we programmed the GeoStoryTelling’s back-end to automatically detect whether the information provided is an image, a video, or an audio, the data gathering process to populate our database follows three main divergent approaches.
1. Audio recordings and Videos need to be embedded, as we show in Figure 4(a) and 4(b), respectively. Video and Audio Embedding, and Image Linkage.
2. Images may be retrieved from their hosting location on the internet, as we show in Figure 4(c).
3. Images can be loaded from our computers, locally, as shown in Figure 4(d).
Although the gathering or compilations of these elements may be tedious with large scale projects, note also that these contextualizing elements are optional for it is only the narratives and the latitude and longitude coordinates that are required for the GeoStoryTelling software to render the HTML outputs.
Once we have collected these audios, videos, or images, we can then populate the column called “Video, Image, or Audio Link” in Table 1. As indicated in Table 1, the video and audio need embedding, which is a specialized HTML code that allows those objects to become part of our GeoStoryTelling maps. However, following our no-code approach to empower researchers or ease access to data science, we can copy that embedding code by selecting “share” from most public streaming websites with videos or audio, as we show in Figure 4(a) and 4(b), and then select the option “Embed.” After doing so as shown in Figure 4(a) and 4(b), we can simply copy and paste those links to our database, as represented in Table 1.
As we indicated, we programmed the back-end procedures of GeoStoryTelling (access to this code here https://cutt.ly/WeFrxG3J) to automatically detect whether we are providing videos, audios, or images, so we do not need three columns, but only one as shown in Table 1.
For the sake of completeness, note that we added one extra row to Panel D in Table 1. This is to illustrate the three main sources of information that the column “Video, Image, or Audio Link” can handle. The first row in Panel D, by starting with a web address, indicates that the output should be an image. The second shows a local directory located in the disk “C.” Note that we do not need to add quotations to any of these formats, we programmed GeoStoryTelling to automatically handle blank spaces. The third row in Panel D starts with <iframe, which suggests that this is an embedded link to either an audio or a video. With respect to images stored locally, please note that, although our example is demonstrated with a Windows path, the same process to add a local image in a Mac computer is executed, also without having to include quotation marks when pasting this path in our databases. If we want to publish our interactive maps online, I recommend to add image links rather than reading them from our computers, for online hosting services may be unable to load those images stored locally.
Finally, the software also allows us to include an external link to a webpage that provides more information, and this link may simply be copied from any web browser address and paste it in our table. Note that in Panel D in Table 1, the first cell in the “Link to Webpage” column is empty. This is completely fine for GeoStoryTelling can handle missing cases if no more information about a given story or narrative is added—nonetheless, stories with complete information will display this external link. As shown in Figure 2, when this link is present, this is added under the name “Hyperlink for more information.” As its name indicates, when users click on this hyperlink, they are taken to a webpage that provides more information about the story. For example, in addition to displaying a short documentary called “Surviving Kensington” in the map via an embedded video, the “Hyperlink for more information” featured in Figure 2, takes users to a TIME Magazine story. 5 Although this story was not written or produced by the same authors as the video documentary featured in our GeoStoryTelling, it may potentially enrich our understanding of the topic and the stories we are sharing.
This concludes our discussion of the creation of the database as shown in the left-hand side of our flow diagram (Figure 1). Once this database is saved as “*.csv” file, it can then be loaded into the GeoStoryTelling user interface, as we discuss next. Although this description may seem overwhelming, workshop offerings with participants with no prior experience in GIS nor computer programming have consistently indicated that, a full completion of datasets with 10 cases were achieved in one hour, including the explanation time of each of these elements. Workshop participants in Europe and South America indicated that the creation of *.csv files in Microsoft Excel separates columns with semi-colons instead of commas. In these cases, the Microsoft setup must be updated to use commas. Alternatively, Google sheets may be used to create these databases and then save them locally as *.csv files before uploading them to the GeoStoryTelling software.
GeoStoryTelling User Interface
Figure 5 shows the GeoStoryTelling’s user interface (UI). As described earlier, this UI comes pre-loaded with a toy dataset that can be used to replicate Figure 2. As illustrated in the flow diagram, we can start by loading a database we have created following our previous discussion. Alternatively, to get familiarized with GeoStoryTelling, we can also select to load the dataset included with the software. To achieve the latter, we can select the option “Minimal example with six cases” under section “A. Load your data or the example” and then select the option “Click to load GeoStory data example shown below” under section B in the UI. If instead you want to use your own data, you can drag your “*.csv file” to the UI or search for your database directly by clicking on “Click to browse from your CSV…” as shown in the GeoStoryTelling UI in Figure 5 also under section B in the UI. GeoStoryTelling User Interface. Mac users can download GeoStoryTelling here https://cutt.ly/E724wAx, Windows users here https://cutt.ly/J724f0C.
Once our data are loaded, such a database will be automatically displayed in the UI. To execute GeoStoryTelling, you just need to select the columns that correspond to the options in section C. In the toy database we are providing, the columns of the database match the column names we discussed above starting on page 18. As indicated in the flow diagram (Figure 1), the only required columns to obtain the interactive maps are the latitude and longitude coordinates. The remaining columns discussed in Table 1 are highly recommended but optional.
Note that GeoStoryTelling was programmed to automatically recognize the column names of our databases loaded to the UI. These column names will also match the database displayed within the GeoStoryTelling’s UI once our databases are loaded. To select these columns in section C, simply click on each of the options we are displaying in the left-hand-side of the UI, as shown in Figure 5, and select the specific column name conceptually matching this option. For example, in the toy dataset we are referring to latitude as “lat,” accordingly, under the option “Latitude” simply click the dropdown menu and select the column “lat.” A similar rationale applies for the rest of the columns in our database.
Finally, note that GeoStoryTelling allows users to provide their own titles for the resulting HTML visualization. The default text for the title (added as a place-holder) reads “Type a short title or delete this text.” This implies that if users do not want to add any title to their resulting maps, they should delete this text in the UI—if they do not delete this text, the map will display “GST Title: Type a short title or delete this text.” GST title is included in the back-end of GeoStoryTelling. When users want to modify the title to be displayed in the HTML map, they can type the title they best believe captures the goal or purpose of their GeoStories.
Once all desired options have been populated, we can proceed to generate the HTML interactive maps by clicking on the tab called “II. Execute GST.” If the latitude and longitude coordinates were selected, GeoStoryTelling will automatically launch the map in a browser. Additionally, GeoStoryTelling was programmed to automatically save the resulting HTML file locally to be distributed or to be hosted on the internet at researchers’ discretion. Note that in addition to a file called “index.html,” researchers will need to also distribute the folder called “lib.” Finally, every time GeoStoryTelling is executed this file and folder will automatically be rewritten.
Toy Database Included in GeoStoryTelling
For illustration purposes, the database included with GeoStoryTelling (see https://cutt.ly/peFeyXEJ) contains six cases. Two of these cases show embedded videos but from different sources. One is retrieved from YouTube and corresponds to a 2016 presentation focused on applications of spatial data modeling to education and social sciences research. In this case the hyperlink that includes more information about this story simply features a link to the university where that presentation took place. Finally, the text is a short reflection of the meaning of that talk for that presenter. The second video was retrieved from vimeo. This video captures the narratives of inhabitants of the Kensington neighborhood in Philadelphia. Considering that the participants featured in this documentary are experiencing addiction to drugs, homelessness, overdoses, as well as injections, we felt compelled to add a trigger warning as part of the name of this story. The link added in this case leads to another report on this same topic conducted in 2012 by TIME magazine, and the text featured in this story is a description of the video documentary as retrieved from vimeo.
The third story features the live recording of a concert that took place in September of 2022. This recording was retrieved from SoundCloud. In this case, the link associated with this story is a website https://www.1001tracklists.com containing all tracks of that concert as separate audios (or videos for the original versions can also be played in YouTube, for example), rather than being mixed together by the artist featured. The text depicted in this story, also simply accounts for the description as provided by SoundCloud.
In following with our goal to showcase the flexibility of GeoStoryTelling, we are also including a short reflection on “home,” as a during- and post-pandemic place where academic work and learning may occur based on technology access and online resources. In going back to our illustrative goal, in this example, we are also demonstrating how images locally stored in our computers may be embedded in our GeoStoryTellings. The hyperlink selected in this case features that participant’s job-related website.
The last two stories show two examples of literacy-related places. One features the Barnes Foundation as a traditional and “well-established” place where culturally-relevant events are realized or exhibited. The image embedded in the map was retrieved from the gallery of images featured by the Barnes Foundation and the hyperlink takes users to its main website. In this example, the text is an attempt to convey how this museum should not necessarily be considered more important than other places where literacy and cultural activities are happening, such as the last example we are presenting.
The last example features a center where counternarratives of traditional notions of art are taking place. This center, although not exclusively, mostly works with undocumented immigrants, from Mexico and south America. As part of their activities, members of this community create different forms of art (music, dance, paintings) and provide spaces for after school activities and literacy-oriented opportunities for adults. The image depicted in this story shows the role that this center provided by serving as a vaccination site, where the (likely undocumented) community members could be vaccinated without fear based on their legal status in the United States. The text featured also highlights that the cultural activities are not limited to immigrants but also reach Americans (mostly academics from neighboring universities) who are investing time and resources in strengthening the scope and presence of this cultural center. Finally, the hyperlink simply leads to the main website of CCATE, which among other things features ticket sales for their artistic events.
Please note that this dataset was primarily built to illustrate the types of multi-media elements that can be integrated with GeoStoryTelling. In our discussion section we will delve more deeply into one of these featured stories to further illustrate how researchers may use GeoStoryTelling focusing on a more serious health and social issue: the opioid crisis.
Applications and Examples
We were purposefully eclectic in the selection of the stories featured in our toy database; mostly to illustrate, on a generic or “meso-level,” the flexibility of featuring narratives that do not necessarily need to be connected. More realistically, however, a study with a clear purpose will be more uniform in the selection of the narratives, which although individually will likely present important variations, will also convey information about unifying themes. That is, if the topic is drug addiction in a given neighborhood, as we discuss next, although individual stories may vary, it is quite likely we will find common ground or structure across the multiplicity of voices and experiences being captured.
In the following lines, please allow us to go back to the discussion of the selection of narratives or stories being featured in our GeoStories. That is, as stated earlier, although GeoStoryTelling is capable of displaying large amounts of text, the featuring of entire transcripts would be both challenging and likely not very useful. Instead, in the same way qualitative researchers are selecting quotes to illustrate their points in a traditional qualitative paper, researchers relying on GeoStoryTelling may also want to select and geocontextualize quote or quotes that will become central findings that illustrate the points they want to communicate. In terms of deciding how to proceed with participant selection, although in our example we presented different participants or cases in a single map, researchers may instead decide to feature one participant across its space-time continuum per map.
With these possibilities in mind, the following discussion presents two non-exhaustive analytic approaches that researchers could implement. In the first approach we illustrate how our analyses may be more precisely contextualized by zooming-in into the specific locations where our participants’ stories took place. In the second analytic approach, we illustrate how following the evolution of our participants’ stories or experiences over time may further strengthen our understandings and lead to the development of plans of action that may have better prospects of success. Note, however, that GeoStoryTelling is flexible in incorporating a myriad of analytic decisions that may likely evolve organically based on the research team’s goals and the topic of study. At the end then, the strategies featured next are not mutually exclusive, instead they can be merged based on our research goals and data collection strategies.
Stories with More Precise Locations
To illustrate this analytic strategy, let us retake what we consider the most relevant example illustrating the analytic power of GeoStoryTelling that we included in our toy database. This example corresponds to the presentation of the Kensington neighborhood in Philadelphia. As depicted in the video documentary, there were multiple participants being featured. Although all these participants were situated in the Kensington neighborhood, their stories felt somehow aggregated for they represented such a neighborhood globally, even though in the video there are specific places featured (e.g., a park, a street corner, a house) that may be “decomposed” to offer more detailed accounts of these participants’ experiences. Accordingly, we could more precisely feature or pinpoint those locations along with specific geostories being featured. In other words, researchers, instead of presenting a video that features multiple stories in an area that includes multiple blocks, may decide to feature individual stories while situating those stories in the specific places (i.e., addresses) where those stories were recorded or gathered.
This process of “zooming-in” (González Canché, 2019) to prove more detail about location and experiences may increase our understandings of those experiences; and, to the extent that we are able to add more context to those narratives via “hot-links,” more nuance in these participants’ stories may also be gained. We must, however, note that this is a vulnerable population and as such we should also have a clear goal in mind if, or when, deciding to feature these more granular levels of detail. These goals may involve the creation of programs or strategies to address their addiction to drugs or provide them with shelter options, or both. We must also be aware of the added risks we are creating for these participants by not only overtly showing their identities, but also revealing their locations. Their addictions may make them target of many types of abuses, even more than before they agreed to be featured in this type of research or documentaries.
Fewer Case Stories Followed Over Time
In addition to decomposing multiple stories, we could also focus on documenting the pathways that brought our participants to the time and place where the interview was happening. For example, assume we may be interested in identifying now inhabitants of Kensington who did not grow up in this neighborhood or even in Philadelphia. Here, the goal may be to understand the processes they experienced to “find Kensington” with an emphasis on highlighting their first experiences in or the points of entry to this neighborhood. That is, we could aim to understand not only what their entry points were, but also the processes and reasons that led them to reach those entry points. In this form of data gathering, the contextualizing tools may be their voices, linked to their places, and the text featured in those locations may be our reflections or our analyses of their experiences. Ideally, we may be able to decompose their stories based on global time-stamps (González Canché, 2022; 2023) where we may reconstruct the temporal evolutions of their stories based on their episodic memories or their recollections of the events that brought them to the time and place where they shared their stories with us.
Far from a voyeuristic endeavor, the goal of documenting and understanding their trajectories may serve to see if there are commonalities or structures explaining entry points into “Kensington.” We are using quotation marks when referring to Kensington here because “Kensington-like” places, also known as “open-air narcotics markets” are unfortunately becoming more prevalent across the United States (Percy, 2018, para. 5). Accordingly, to the extent that we are able to understand entry points, their mechanisms, and the types of structures supporting their prevalence, plans of action may be pre-emptively designed to offer support to future “Kensington inhabitants or newcomers” before reaching homelessness or more acute drug addiction problems. This is important because “drug tourists who come [to Kensington-like places] may never leave” (Percy, 2018, para. 7).
Application Example to Survey Responses Involving Campus Safety
In this section we are providing an illustration of how GeoStoryTelling may be used to gather data via online surveys that may be larger in scope. In this example we are requesting three prompts designed to protect the identity and preserve the anonymity of the survey respondents. The following lines describe how we recommend phrasing the instructions. Language related to Institutional Review Boards in the United States and Canada and the General Data Protection Regulation in Europe should be included in actual implementations of this example.
Proposed Instructions
With the goal of strengthen our efforts to provide you with the best possible on campus experience, we are asking your support to share with us any experience you may have while on campus where you felt unsafe, threatened, or unwelcomed, while briefly elaborating on the issue or issues that made you feel that way.
All your responses will be completely anonymous for no names are asked and your participation is voluntary. However, in order to gain as much nuanced understandings, we would appreciate if you can provide us with your pronouns and ethnicity if you feel comfortable sharing this information with us. Additionally, we are interested in mapping the locations where these incidents took place, which is why we ask you to describe these locations as detailed as possible, like if you were to look for these places using Google Maps. You can be specific, for example, in the case below the respondent mentioned second floor of Stiteler Hall at The University of Pennsylvania, which in Google Maps pointed to the Graduate School of Education student records office.
The specific fields we are requesting information are the following. Note that each prompt also contains example responses below:
In this example we can see how narratives may be collected and even if these responses amount to hundreds or thousands of cases, the process of geotagging them is straightforward. That is, once our initial database has been downloaded from services like Qualtrics or SurveyMonkey, for example, we can copy the responses of the place of incident field (prompt or question number two in our example) and paste it in GPS Visualizer
(https://www.gpsvisualizer.com/geocoder/) as we illustrated in Figure 3. As discussed above, this batch geocoding method allows the effortless retrieval of thousands of latitude and longitude coordinates with free services like Bing Maps—note that we also mentioned that we can create a free API key above to access Bing map servers at https://www.bingmapsportal.com. Please note that this example is for illustrative purposes, its implementation would require Institutional Review Board or General Data Protection Regulation approval. Also note that, although no names are requested, there is the possibility that the narratives do include names of alleged aggressors. Accordingly, the research team must be extremely careful in vetting and curating the data to avoid any mishandling of information or unfairly damaging the reputation of individuals involved in these incidents.
In addition to geomapping the raw responses as provided in the original incident descriptions (third question or prompt shown above), in data collections where these responses reach hundreds or thousands of cases, researchers may also apply machine learning text classifications to label these incidents into categories. This process may be useful when trying to label and synthetize or summarize these incidents. This process may be achieved with another no-code software available at Expert Systems with Applications (González Canché, 2023) or the International Journal of Qualitative Methods (González Canché, 2023). The output of these text classification analyses will have the same information in terms of location that can be geocoded using GPS Visualizer as depicted above. The main difference is that we now may map these incidents based on a count of similar classes of issues, namely threats or discrimination, rather than listing the original comments. The decision to present original responses or the classes of incidents, depends on the goals of the research team and in all instances, we must prioritize fairness and protect the identity of all parties involved.
Limitations and Future Steps
We understand that the opioid crisis and harassment or discrimination issues represent important social and health related problems that are more dire than many other research topics in the social sciences. In a sense then, by illustrating how GeoStoryTelling may be used to provide more nuanced understandings based on geocontextualizing these complex cases, we are also demonstrating the salience and flexibility of this methodology for social and health scientists. To illustrate this analytical flexibility, we also provided an example of how researchers may apply GeoStoryTelling to survey responses, either to present “raw stories” or even to first classify those stories and then map the resulting categories.
Despite these illustrations, we, however, are not claiming that our proposed implementations and the operationalization of our methodology is without any flaws. In addition to the added risk that disclosing locations, and potentially but not necessarily, the identities of our participants (we can use pseudonyms or blur their faces, for example), there is also the issue of selecting the specific narratives we decide to feature in our geostories. As indicated above, whenever possible we should seek the advice of our participants in validating whether our narratives or stories selection actually captured their experiences—via interactive reviewability (Alexander et al., 2019; González Canché, 2022, 2023).
Relatedly, whenever possible we should add, as part of the narratives we are selecting (i.e., the actual text included in the column “Geo-Story” as shown in Table 1), place-based indicators that we, or our participants, may have considered worth mentioning to provide more nuance to their resulting geostories. Specifically, we could add, as part of their text narratives, attributes of the locations where those stories took place that may help us better understand their stories. This process goes back to our initial discussion of the relevance of the neighborhoods where we experienced life in influencing our decisions, fears, beliefs, and perceptions of the world (González Canché, 2023, 2023; González Canché et al., 2025). Although these place-based indicators could be mapped using GIS methods and data sources, there are two main concerns that deterred us from including those attributes in the current version of GeoStoryTelling.
First Main Concern
The first concern is that data collection would be much more complicated than the one we illustrated here—which will require some form of GIS training, and at least some basic statistical programming, thus defeating the goal of GeoStoryTelling that aims to avoid both of these barriers, as we further discuss below—see González Canché (2023) for a set of low-code tools to apply GIS in social science research. That is, in the easiest approach 6 the data collection will require accessing place-based indicators from the American Community Survey in the United States, for example, (or other agencies in other countries) along with the shapefiles (more on these files below) required to actually map one of these indicators at the time (González Canché, 2023). That is, visualization of indicators via shapefiles handles one indicator per map, unless several layers of data are added to the resulting HTML visualization (González Canché, 2023a, 2023b, González Canché et al., 2025a, 2025b; González Canché & Zhang, 2025).
Let us briefly discuss shapefiles. First, note that based on openstreemaps.org, GeoStoryTelling includes points, lines, and polygons (González Canché, 2023), to represent buildings (i.e., points or the intersection of a latitude and longitude coordinate), streets or rivers (i.e., lines or the continuous intersection of a collection of points), and blocks, zip code tabulated areas, cities (i.e., polygons or the intersection of lines that end at its starting point). What GeoStoryTelling is not accounting for is place-based attributes that may be merged with those shapes to add color, for example, to the resulting points, lines, or polygons.
Even though the process of merging shapefiles with place-based attributes is not complicated, per se (González Canché, 2022, 2023a, 2023b; González Canché et al., 2025), this merging process requires feature engineering procedures (i.e., advanced data cleaning and preparation) that may also require at least some form of statistical programming proficiency and/or access to proprietary software like ArcGis. From this view then, the addition of place-based features that require their merging with shapefiles, in essence defeats two of the main goals of GeoStoryTelling. The first is ease of use, that is, we designed GeoStoryTelling to expand access to data science and visualization tools that do not require any computer coding or statistical programming. The second is that we are offering GeoStoryTelling free of charge so that cost is not a barrier toward its utilization.
Second Main Concern
On a more substantive level (i.e., moving beyond technical expertise and monetary costs), depending on the episodic memories of our participants and considering the non-linear temporal dimensions in ethnographic and qualitative research (Braun & Clarke, 2006), ideas shared by our participants may prompt them to remember other past events that may have shaped their experiences. In these cases then, the inclusion of place-based attributes across multiple time spans or time periods, is not currently feasible to achieve even with the newest developments in spatial visualization (González Canché, 2023). Rather, place-based indicators are currently constrained to show a snapshot in time of how certain areas were expected to fare. This expectation is provided by surveys like the American Community Survey in the form of “estimates” of place-based indicators like unemployment or household composition, for example. That is, these estimates are statistical predictions of those place-based attributes, as opposed to the actual measures of such indicators (U.S. Department of Commerce, Economics and Statistics Administration, U.S. Census Bureau, 2020) (hence the relevance of considering margins of errors when utilizing these estimates). Having said this, there are studies that benefit from the inclusion of these indicators. For example, González Canché et al. (2025) added indicators of place-based poverty to identify cases of high achieving low-income students (see https://cutt.ly/hwvraSQD). Although GeoStoryTelling does not currently incorporate this possibility. A given place-based indicator may be added to our geostories or geonarratives, thus still conveying the geo-contextualizing information that may enrich our understandings—as we illustrate further next.
How to Address These Apparent Limitations?
Although we decided not to include traditional place-based estimates in the GeoStoryTelling interactive maps, note that the texts, pictures, videos, sounds, and links to other sites we are adding to each point (i.e., intersection of latitudes and longitudes), are indeed qualitative attributes that help us understand that point. From this perspective, our maps do contain attributes and these attributes are not statistical predictions or estimates, but more precise accounts of our participants’ experiences.
Nonetheless, we can also more organically address the absence of traditional “quantitative” estimates across time associated with those locations, by purposefully being more strategic during the data collection process. That is, if we aimed to geocontextualize our participants’ experiences in issues related to place-based safety, for example, we should prompt them to include place-based details that may help us better understand their experiences. In returning to one of our previous examples, which featured the experiences of one participant who attended graduate school in Los Angeles California, as researchers, we could follow up our conversation with them as follows:
In this example we were able to gather two key pieces of information that were space and place (splace) related. In terms of space, as we discussed earlier, its meaning may be more abstract and, in this case, it may imply notions of distances. For this participant’s experience, commuting time or distance was quite uncomfortable not only for the distance itself, but also for the type of neighborhood where her apartment was located. Specifically, and in terms of places, it seems that she not only did not like her neighborhood, but also she felt quite unsafe, to the extent that she preferred to spend the night at the library rather than to risk going home after dark, or arriving home after dark. These two pieces of information should be added to the resulting text narratives selected to be featured in the GeoStoryTelling visualizations for they will help us to better understand the source(s) of her challenges during graduate school. Moreover, these details will also address the issue of not including other quantitative place-based estimates for the reasons mentioned above.
This example illustrates what Ajayakumar et al. (2019) were referring to, when describing the relevance of including context that goes well beyond merely adding layers with quantitative estimates of sociodemographic and socioeconomic place-based indicators. In their words, the interest is not only in indicating that something occurred, but more importantly on explaining why it did, or how it did (Ajayakumar et al., 2019).
Once more, we should double check with our participants that our interpretations are accurate and overall try to give them as much agency as possible in the sharing of their stories. After all, GeoStoryTelling as a methodology is only relevant to the extent we are accurately capturing the geocontextualized experiences of our participants.
Is GeoStoryTelling Just Another Software?
It is worth noting that some of the features contained in the output rendered by the GeoStoryTelling software can be approximated with Google Earth. In this respect it is worth noting that, like our software, Google Earth cannot incorporate layers of quantitative data. Unlike our software, Google Earth cannot support audio or video in the resulting visualizations. Having said this, Google earth can measure distances and generate paths, both of which our GeoStoryTelling software cannot yet accommodate. If the latter are important elements for the geostories to be shared, then we recommend users to apply the methodological framework we are presenting here and create their geostories with Google Earth, even if this platform does not enable users to upload their data, like our GeoStoryTelling software does. In this respect, note that we are not arguing that the software accompanying our framework is better than any other software available on the market. The true values added of our software are that it is completely code free, free to use, and free to distribute, and runs locally. This means, we can obtain state of the art interactive visualizations without any coding or programming experience, and we do not need to upload any of our data to any external server. More importantly, we are providing this software as a complement to the methodological framework presented in this paper. That is, GeoStoryTelling it is not just another software for users may still apply the concepts we are describing in this study and execute them in other software tools that may even require computer and statistical programming proficiency.
Potential Learning Curve
Despite being user-friendly, researchers might still require time to familiarize themselves with the software’s functionalities, especially in integrating various multimedia elements. This is why in this subsection we are offering an estimate of the time needed to create a database and becoming fully familiar with the software capabilities based on workshop offerings wherein participants have started from no prior knowledge and have achieved the expected HTML outputs using the GeoStoryTelling software tool.
We have had the opportunity to offer free professional development workshops to undergraduate college students at the University of Guadalajara in Jalisco, Mexico and to high school students in Singapore, and Shanghai and Guangzhou in China. In each of these workshops, we spent about one hour in total with a total of 120 students (about 25 in each workshop offering—with five workshop offerings in total) to create usable databases that rendered the expected interactive HTML outputs. Note that this time included the explanation from no prior knowledge of the software rationale and capabilities as well as the explanation of the processes and steps needed to gather data, retrieve latitude and longitude coordinates, and compile links to connect the GeoStories to other websites. This does not mean that complete research projects may be achieved in this amount of time, the exercises consisted of asking participants to identify splaces (places and spaces) that were somehow meaningful to them and then create the respective datasets that incorporated images, texts, hypelinks, and video or audio sources. It is worth noting that these participants are assumed to be tech savvy given that they grew up with technological tools. Older trainees who may not necessarily have grown with technological tools may require longer times to build their datasets. Having said this, although not in workshop contexts, we have individually trained qualitative researchers in their sixties, without any issue. In these few cases (three in total) the resulting training time consisted of two-hour sessions, but, once more note, that these were individual sessions, as opposed to groupal training sessions.
Finally, please note that as part of our ongoing collaboration with ethnographers, qualitative, and mixed methods researchers, we have started collaborating with researchers whose main form of data collection is through stories. Some of these ethnographers work with third and fourth grade students and they commented that they have started to follow the guidelines depicted in the paper to geocontextualize the stories generated by their students. They have also mentioned their intentions to teach their students how to build their own datasets to use with the GeoStoryTelling software. In this respect, these researchers are confident that their students have enough technological familiarity to generate and understand the steps required to compile these data repositories and datasets.
Technical Support and Sustainability
As part of our efforts to ease access to GIS and interactive visualizations, we are committed to offering technical support and maintaining the software updated. In terms of technical support, we will continue to offer yearly professional development workshops in international conferences like the American Sociological Association and the American Educational Research Association. More to the point, in addition to providing our contact information in the user interface, we are in the process of recording tutorial and instructional videos focused on troubleshooting common end-user issues that may emerge during the database creation and curation, especially when working with large scale research projects.
With respect to software updating and sustainability, we are committed to maintaining the software updated in both Code Ocean and in the Windows and Mac versions, in all instances we will document all changes made. For example, due to high costs, Stamen (see https://maps.stamen.com) was unable to continue providing its base tiles, which were used by GeoStoryTelling. Although this change impacted the rendering of the HTML visualizations in version 1.0 of GeoStoryTelling, the solving of this issue simply required the changing of the tile provider which Code Ocean reviewers accepted and it is now published as version three, see https://codeocean.com/capsule/6296897/tree/v3.
Final Thoughts
The only requirement for the GeoStoryTelling software to work is that each of the selected texts or narratives need to be linked to the spaces and places (splaces, González Canché, 2023) we want to incorporate in our analyses. This process requires us to have a clear understanding of where we need or want to situate these narratives. Even in the Kensington neighborhood, where there exists an overt and clearly defined area, the selection of specific splaces to be featured in our GeoStories may include an aggregated depiction of the neighborhood or may instead share more precise locations. In other instances, like with interview data, for example, the selection of splaces will depend on the episodic memories of our participants. For instance, we may be in Spain during some aspects of the story and in Los Angeles during other time periods. In both cases, we strongly recommend consulting with our participants to not only reach common understandings of where the narrative should be situated but also to validate our interpretations and analytic decisions—i.e., interactive reviewability (González Canché, 2022; 2023).
As mentioned above, our GeoStories may need to feature the same participant in multiple splaces and across different points in time. If this is the case, our database construction should reflect this splace-time continuum, and the ID should reflect this time that will become part of the resulting interactive map’s legend. Recall also that the measure of time is relative to the specific experiences shared by our participants. That is, our reliance on global time stamps provide us with flexibility in the selection of units of time. While for some participants in the same study, time may be measured in years or decades, for others, time measurement may be operationalized in days or weeks. Both inputs should be respected and captured as precisely as possible. We hope that we made clear that GeoStoryTelling is capable of capturing these global and non-linear time stamps and that the process to build the resulting databases to be loaded in the user interface may be achieved without any need of specialized external software. This easiness of the analyses, however, does not mean that the database construction is not time consuming, for, depending on whether we want to add hot links, images, video or audio, the compilation of these optional elements may be time costly. If, instead, we are satisfied with only featuring narratives that are geo-contextualized, like in our survey example, then the dataset construction is less time demanding, but the resulting geostories would be less contextualized, nuanced or rich. In short, researchers may need to decide what elements they are interested in featuring when building their stories.
We do hope that researchers in training and professionals alike may interact with the GeoStoryTelling user interface, either by building their own datasets or at the very least by interacting with the toy database we are providing. We are convinced that access to these tools may lead to enriching stories to be shared and hopefully to the discovering of structures and mechanisms that may aid in the development of effective plans of action. The time has come to remove barriers that continue to prevent ethnographers, qualitative, and mixed methods researchers from benefiting from the analytic and structure unveiling capabilities that sophisticated data science and visualization tools bring to social and health sciences research. Finally, note that GeoStoryTelling may also be used for investigative reporting like those stories features in the Washington Post or the New York Time, thus going beyond academic research endeavors.
Footnotes
Acknowledgement
I am grateful for all the inputs and feedback provided by the reviewers and editors.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the University of Pennsylvania' Penn Global, the Spencer Foundation, and the National Academy of Education. The findings and contributions of this study do not represent any of those funding agencies and all errors remain those of the author’s.
