Sage Journals: Discover world-class research

Abstract

This study aims to evaluate 16,891 academic publications in the field of cinema between 1980 and 2024 using bibliometric analysis and topic modeling methods. Based on data obtained from the Web of Science (WOS) and Scopus databases, bibliometric findings were received, including the distribution of publications by year, the annual number and rate of citations per article, the most productive authors in the field, the production status of authors over time, the countries of authors and the number of articles they published, and the journals with the highest number of publications. Data obtained from the Web of Science (WOS) and Scopus databases were also used to identify prominent word groups and themes in the articles using text mining and Latent Dirichlet Allocation (LDA) topic modeling. As a result of the analysis, 12 main themes emerged based on word-text relationships and the weight of publications. The findings show that cinema studies have developed with increasing momentum over the years and that there has been a growing focus on certain topics. This study systematically examines the development of cinema studies literature through descriptive content analysis and LDA topic modeling. In this respect, it is important in that it systematically reveals the structural and thematic transformation of academic production in the field of cinema and provides a theoretical and methodological basis for future research. It also makes a current and multidimensional contribution to the discipline in terms of revealing the increasingly important digital trends, cultural representations, and interdisciplinary developments in cinema studies.

Plain Language Summary

Using Machine Learning to Identify Future Trends in Cinema Through Thematic Mapping

Keywords

machine learning text mining LDA bibliometric analysis cinema film studies motion pictures thematic mapping cinema phenomenon future trends

Introductıon

Conceptually, cinema is a branch of fine arts derived from the French word “cinematographe” and possesses unique characteristics. In this sense, cinema, which has been accepted as the seventh art form, drawing inspiration from existing art forms such as sculpture, painting, photography, and theater, has become a kind of integrator of all the art forms it has drawn from and a unique medium of expression. However, cinema should not be viewed solely as the result of an artistic approach, but also as a means of communication, a cultural representation, and a producer of political discourse. In this context, the perspectives presented by cinema theories explain the multifaceted and multi-layered structure of cinema. Roberto Rossellini defines cinema as the manipulation of reality through images and sound (Michelson, 1995), while Bazin (1967) describes it as the capture and reproduction of a segment of time, focusing on cinema’s relationship with reality rather than its artistic aspect.

Discussions on the nature of representation within cinema theories have intensified, especially after the 1970s. Christian Metz’s semiotic approach analyzes cinema as a language system, while Laura Mulvey’s feminist film criticism questions the male gaze on the cinema screen and patriarchal regimes of representation (Mulvey, 1975). As can be understood from these approaches, it is generally accepted that cinema does not merely produce content, but also possesses a sign system with ideological functions.

Cinema is not only Western-centric in terms of its industrialized structure. It is also largely guided by Western-centric theoretical paradigms in terms of its academic interpretation and theoretical organization. Although this situation began to be questioned with the rise of postcolonial film criticism, the views that emerged after Said’s (1978) conceptualization of Orientalism brought to the fore the debate that cinema functioned as an instrument of cultural hegemony in the postcolonial period. The works of Bhabha (1994) and Spivak (1988) have attempted to explain how the representation of the other in cinema operates in the construction of identities.

Cinema has come to the fore as a fictional space that has continuously updated itself with new formats added over more than a century and strengthened its technical equipment to increase audience participation in the content. Cinema, which has developed in many ways, has made its most striking updates in the elements that are prioritized within hierarchical perception practices. From the first film screening in 1895 to the present day, cinema has become an innovation that the whole world cares about with its effective position on the global stage. This situation has made it necessary to re-examine and re-evaluate cinema in the context of technological transformations such as digitalization, media convergence, and artificial intelligence, beyond classical theories. In this context, the integration of technology-based developments into cinema has brought about changes in both the definition and functional processes of cinema and in the viewer experience. Manovich (2001) evaluates digital cinema in the context of software-based cultural production, while Hansen (2004) argues that digital cinema has transformed the viewer experience in terms of body politics.

Cinema is a multifaceted art form. It encompasses not only art, but also economics, ideology, culture, communication, and all disciplines centered on humanity, making it a development that attracts the attention of all segments of society. Therefore, it is only natural that many disciplines and academic circles interested in cinema engage in discussions and scientific research with cinema at the center. This multifaceted development of cinema has led to diversification not only in film production but also in the ways it is interpreted. Cinema, which attracts the interest of different disciplines such as art, ideology, culture, economics, and communication, has begun to take center stage in interdisciplinary studies. Therefore, there is a need for cinema studies to be evaluated in a more systematic, data-driven, and multidimensional manner today. In this sense, the multi-layered and multifaceted nature of cinema and the increasing interest of many disciplines in cinema have highlighted the need for a bibliometric and modeling analysis of cinema. The study systematically analyzes the development of literature in the field of cinema through descriptive content analysis and LDA topic modeling. In addition to being based on the current general findings, this study was inspired by previous bibliometric studies and aims to highlight important gaps for further research.

In recent years, artificial intelligence and natural language processing techniques have begun to be integrated into cultural fields such as cinema studies. Modeling techniques such as LDA (Latent Dirichlet Allocation) have become important tools for extracting topics from textual data. This study aims to reveal both thematic and structural trends by examining articles published in the field of cinema between 1980 and 2024 using bibliometric analysis and LDA-based topic modeling methods.

Main Problems of the Study

The main problem of this study is to determine which themes have come to the fore in academic publications in the field of cinema between 1980 and 2024 and to develop predictions about future research directions in line with these thematic trends. Cinema is not only an artistic form of expression; it is also a multidimensional field of communication that reflects cultural, ideological, and technological transformations. Therefore, identifying which topics are more prominent in cinema studies is important for understanding the dynamics of development in the field. However, systematic, data-driven, and large-scale analyses such as bibliometric analysis and topic modeling on cinema studies are quite limited. This situation highlights the need for a comprehensive assessment of the structural and content-related transformation of academic production in the field of cinema.

The findings obtained within the scope of the study have been interpreted in the context of theoretical inferences about the historical course of academic production and future orientations, within the framework of structuralist theory and cultural transformation theories. This theoretical grounding is based on the assumption that the field of cinema has been reshaped along the main axes of socio-cultural transformation, digitalization, political representation, and interdisciplinary expansion.

In this context, the study sought answers to the following questions:

What themes does scientific production in the field of cinema focus on?

How have these themes changed over the years?

What topics are cinema studies likely to focus on in the future?

Theoretical Foundation

The phenomena emerging within social organization are grounded in a set of structural and semantic relationships. Structuralist theory is concerned with these underlying structures. Structuralism, which has been frequently used in the analysis of language, culture, and society since the 20th century, does not seek to explain everything. Instead, it attempts to reveal the structure of the work by asking real questions (Furtana, 2014). In cinema research, works are not merely the sum of their contents but gain meaning within specific linguistic, cultural, ideological, and social structures. Therefore, themes derived from bibliometric distributions and topic modeling analyses are shaped by the influence of certain structural codes (e.g., gender, political representation, cultural identity). The findings obtained during the study reveal the historical and structural relationships of cinema studies. Uncovering the structural nature of dominant themes in the past and proposing ideas for the future is based not only on statistical but also on structural analysis-based approaches.

In the modeling and trend analysis process of the study, future trend predictions were also interpreted within the framework of cultural transformation theory. “Cultural transformation is a result of cultural interaction emerging in a globalizing world” (Özekici & Ünlüönen, 2019). Cultural transformation theory is primarily concerned with changes in cultural practices and representations. In particular, structural transformations such as technological developments, digitalization, and globalization have led to the reshaping of cultural practices and representations; this situation has necessitated the examination of these transformations within a broad academic context, and the phenomenon of cultural transformation has become a common area of interest that brings together many disciplines around the axis of structural change. Themes that are understood through modeling and gain momentum through trend analysis show that the field of cinema is undergoing not only technological but also ideological and cultural transformations.

Purpose and Significance of the Study

The main purpose of this study is to examine academic publications in the field of cinema between 1980 and 2024 using bibliometric analysis and topic modeling methods, to identify the dominant thematic orientations in the field, and to provide guiding insights for future research. The study also aims to examine how cinema-related research is approached in different disciplines, revealing interdisciplinary transitions and the diversity of theoretical approaches.

This research is significant in that it is conducted with a large data set of 16,891 articles. The sample size is a fundamental element in bibliometric analysis that enhances the reliability and generalizability of the results obtained. Additionally, the LDA topic modeling method used in the study objectively and data-driven identifies prominent themes in the field of cinema, thereby contributing to the literature both methodologically and substantively.

Method

This study presents scientific results using bibliometric analysis and topic modeling analysis methods consisting of data collection, inclusion-exclusion classification, text preprocessing, topic modeling, revealing ideal topic numbers, and trend analysis for the concept of cinema, which has been examined by a wide range of disciplines in the scientific literature.

Academic articles were used as the unit of analysis in this study. Each article was evaluated at the level of title, abstract, and keywords, and topic modeling was performed based on these contents. Research themes represent semantic clusters determined through the LDA algorithm. In addition, publication year information was also used to enable the analysis of temporal trends. Thus, content, conceptual, and temporal analysis units were evaluated together.

Type of Research

This study aimed to reveal the bibliometric profiles of cinema-related studies and identify important topics as a result of topic modeling. In terms of type, the research falls under the category of descriptive content analysis. Descriptive content analysis, a method within content analysis, involves systematic studies that evaluate trends and research results in an explanatory dimension by including all published studies related to a predefined topic in the data set (Lin et al., 2014; Suri & Clarke, 2009; Ültay et al., 2021). The results obtained from this analysis method are expected to guide future studies on the targeted topics (Lune & Berg, 2017; Yıldırım & Şimşek, 2018). Due to the use of descriptive content analysis in different studies, it is necessary to present the literature in a systematic manner (Ültay et al., 2021).

Application Steps

The analysis process in this study was carried out through the following steps.

Data collection and scope determination,

Data cleaning and application of inclusion criteria,

Text preprocessing

LDA topic modeling and determining the ideal number of themes,

Coding and thematic naming process,

Determining temporal trends with trend analysis

Collection of Data

The data for the study were obtained from both the WOS and Scopus databases using key indexes. Some limitations were applied to the search criteria among the millions of data in these two databases. In the data collection process, where the concept of “cinema” was used as a keyword, book chapters, early-appearing articles, conference abstracts, and conference full texts were excluded, and only published articles were included. Additionally, articles related to the subject written in Spanish, French, and Portuguese, in addition to English, are available in the WOS and Scopus databases. Articles written in languages other than English were also excluded from the study. The most important reason for excluding articles written in languages other than English is the limited access to the literature on cinema studies in other languages and the technical limitations of programs such as RStudio and Orange used in the analysis in terms of linguistic diversity. Therefore, considering the possibility that some words and phrases in many languages that are difficult to understand globally could mislead the analysis tools and expert researchers in the interpretation phase of the study, it was decided to exclude languages other than English from the scope. In summary, this is due to the limited accuracy and consistency of machine learning-based methods such as text preprocessing and LDA modeling in languages other than English. However, it is believed that this shortcoming can be addressed in future studies by using multilingual text mining models (such as multilingual BERT and LDA2Vec).

In this study, articles related to cinema published in the WOS and Scopus databases were also subject to historical time constraints. In this context, articles published before January 1, 1980, and after December 31, 2024, were excluded from the scope.

When examining Figure 1, the quantitative status of the data obtained from the search using the key index is shown in the PRISMA flow diagram. Between January 1, 1980, and January 31, 2024, a search using the keyword “cinema” in the WOS and Scopus databases yielded 22,061 articles in WOS and 31,358 articles in Scopus.

Figure 1.

Collection of research data according to the PRISMA flow chart.

Since the study was limited to published articles, a total of 20,451 arguments were excluded from the study, with 3,282 from the WOS database and 17,169 from the Scopus database. Additionally, 16,077 articles were excluded from the data pool as they were present in both databases. The data were independently reviewed by the researchers, and a consensus was reached on those that met the inclusion criteria in order to minimize bias. After all these processes, the analysis phase proceeded with 16,891 articles.

Analysis of Data

Three different analyses were used in the study.

Bibliometric Analysis

Bibliometric analysis has become a popular method in recent years due to its benefits in processing large volumes of scientific data and creating a high research impact. In particular, bibliometric analyses are preferred in scientific studies for reasons such as revealing trends in article or journal performance, collaboration models, and research components (Donthu et al, 2021). This method reveals the changes and developments in the research topic over time with concrete data. The large and objective data included in bibliometric analysis can be interpreted with subjective evaluation. Bibliometric analysis includes two basic approaches: performance analysis and science mapping. Performance analysis involves evaluating the impact of researchers, institutions, and countries using criteria such as total publications, author contributions, and citation indicators, while scientific mapping helps to map the structure and dynamics of research (Analysis of cited works to identify the most influential publications, co-citation analysis to better understand the relationships between cited works, bibliographic links to connect related publications, co-word analysis to show relationships between topics, and co-authorship to understand the relationships between authors’ social interactions; Passas, 2024). In this study, bibliometric analysis was conducted using the “biblioshiny” application within the Bibliometrix package in RStudio, and the findings were supported with visualizations.

Some inclusion criteria were determined while creating the data set for this study. These criteria were that the document type should consist of articles, the publication languages should be limited to English, and the publications should be between 1980 and 2024. After this filtering process, 16,891 articles were downloaded in CSV format and analyzed using RStudio software. The tools and parameters used in the bibliometric analysis phase are as follows:

Bibliometrix and Biblioshiny packages running in the RStudio environment

Number of publications by year

Authors, journals, institutions, and countries with the most publications

Number of citations and H-index

Keyword analysis

The obtained data were supported by graphs and tables using R-based visualization tools (e.g., ggplot2, graph functions within Bibliometrix).

LDA Topic Modeling Analysis

In the topic modeling phase, which is another part of the data analysis phase, text mining was used. The primary objective of text mining is to classify, group, and label texts. As an extension of data mining based on textual data, text mining is a technology that seeks to derive meaningful results from unstructured textual data (He et al., 2013). Although text mining is performed on unstructured texts, making these texts suitable for qualitative and quantitative analysis is another important goal of text mining. In other words, extracting information and patterns from text documents is the basic function of text mining (Bach et al., 2019).

Text mining methods can be used for many purposes. Classification of expressions and topics, clustering, named entity recognition, sentiment analysis, face analysis, keyword extraction, parsing, topic detection, and topic modeling are just a few of them. Topic modeling is a widely used intelligent technique for topic discovery in natural language processing and semantic mining from documents consisting of scattered, unordered large datasets (Celodar et al, 2018). In the process of the topic modeling method, LDA (Latent Dirichlet Allocation) is a frequently preferred natural language processing method in natural language processing fields. LDA, which provides a productive approach to finding hidden structures in huge information, is used as a heuristic approach to calculate the similarity between source files and obtain the relevant distributions of each document on topics (Celodar et al, 2018). To perform LDA, the data must be converted into a corpus. Then, the compilation text preprocessing stage must be performed.

Text Preprocessing Phase

To perform topic modeling, the data set must first be prepared in a manner suitable for modeling. Therefore, misleading, incomplete, and insufficient data must be removed from the data set to ensure accurate and reliable modeling. The comprehensive examination of the data set, that is, the text preprocessing process, consists of several stages.

Transformation: The first step in the data preprocessing stage is to ensure that all data is included in the same language set. In this context, accents are removed and capital letters in the text are converted to lowercase letters. In addition, HTML extensions in the data set are separated or removed at this stage.

Tokenization: In this stage, the necessary adjustments at the sentence and word level are determined. Punctuation marks and spaces between words are processed in this stage.

Normalization: This stage involves technical processes to determine what type of modeling is required for the data set to be used for modeling.

Filtering: The available data set may not always be organized as the researcher desires. When modeling, both unwanted unnecessary letters and non-English characters can lead to misleading modeling results. At this stage, characters that are not wanted in the modeling, or characters that have no semantic weight, such as “a, an, the,” are removed. At the same time, numbers in the text are separated, and unnecessary punctuation marks and symbols that may hinder analysis are removed.

In the Transformation stage of the study, non-English characters in the text were first removed. Then, to prevent deviations that could occur at the letter level, all letters were converted to lowercase. Again, in this stage, special characters and HTML tags with a high potential to be recognized as words by the model were removed from the text. In the second stage, Tokenization, separations, and arrangements that could mislead the model at both the word and sentence levels were made. Standardization of spaces between sentences and removal of structures classified as undefined words, except for special characters and HTML tags that could be tokens, was performed at this stage. In addition, the texts were parsed at the word level and then lemmatized (films → film, studies → study). In the normalization stage, the words in the texts were reduced to their roots. For example, words such as “films,”“filmmaking,” and “filmmakers” were normalized to a single root, “film.” This process was applied to prevent word variations from disrupting thematic clustering. In the Filtering stage, the final stage of the text preprocessing process, a stopword list was prepared. This stopword list included letters and words that had no semantic weight but had the potential to be tokens. In this study, the words “a, an, the, is, with” were added to this stepword along with all letters. After all these processes, the texts were cleaned structurally and semantically and converted into a vector structure consisting of 1,961,116 tokens. Thus, a high-quality, noise-free text set was obtained for the LDA model.

Topic Modeling and Determining the Ideal Topic Number

In this study, topic modeling was employed to objectively identify the thematic density within the field of cinema. After the text preprocessing stage, the qualitative data obtained were converted into quantitative data using the Latent Dirichlet Allocation (LDA) natural language processing method. In this conversion process, words represent vector structures, and each word is modeled as a vector structure. As a result, a term-document matrix was obtained from the entire corpus.

The topic modeling process was performed using the Orange Data Mining v3.34.1 program written in Python. Orange Data Mining is an open-source data mining and machine learning software developed in the Python programming language with a visual programming interface. It is widely used in the social sciences, particularly for text mining, classification, clustering, and topic modeling, thanks to its modules that support these processes. Through the Text Mining and Topic Modeling modules used in the study, vector representations of the texts were created, the distribution of topics was analyzed using the LDA algorithm, and the outputs were visualized.

The LDA model was applied through embedded libraries, and a certain number of articles under each topic were clustered according to their common keywords. In interpreting the model outputs, the representative keywords of each topic were revealed; the titles, abstracts, and keywords of the articles belonging to these topics were examined through content analysis, and topic names were assigned. In addition, the opinions of two expert academics in the fields of communication studies and cinema were consulted during the naming process, thereby increasing the scientific validity of the interpretations. In particular, content that did not correspond to high-representative keywords or served multiple themes was carefully reclassified during the qualitative analysis process. At this stage, the 16,891 articles in the dataset were evaluated not only based on automatic model outputs but also through the academic orientations, cultural frameworks, and research focuses they represented in the content context. Thus, the topic modeling process was enriched not only with algorithmic depth but also with interpretive and conceptual depth. This method strengthens the mixed methods nature of topic modeling, ensuring that qualitative and quantitative analysis results mutually support each other. The steps of this process are shown in Figure 2.

Figure 2.

Steps of topic modeling analysis.

In the data set collected for the study, each document consists of a combination of multiple topics. In addition, the words used for these independent and different topics contribute to the emergence of certain terms. Based on these assumptions, the most critical element in topic modeling is determining the ideal number of topics to be extracted from this independent and mixed data. To do this, the expertise of the researcher who prepared the dataset is of great importance (Bystrov et al., 2023). Therefore, determining the number of topics in topic modeling is a subjective process. To minimize this subjectivity, there are evaluation criteria for determining the ideal topic in LDA models. In this study, coherence and log perplexity measurements are the evaluation measurements used to minimize subjectivity. After calculating the coherence and log perplexity values, the appropriate number of topics was identified before the topic modeling analysis. By examining the suitability of the coherence and log perplexity values, the ideal number of topics was increased to 30. The coherence and log perplexity values of these 30 topics were compared with each other.

Table 1 shows the 30 topics examined to determine the ideal number of topics, ranked according to their Log Perplexity and Coherence values. The highest coherence and lowest perplexity values indicate that there could be as few as five topics. However, the large amount of data and the token count of 1,961,116 provided a valid reason to increase the number of topics, and after examining the word clusters, it was decided to set the number of topics to 12. The evaluation conducted in the range of 1 ≤ k ≤ 30 to determine the most appropriate number of topics also pointed to 12 topics based on two main reasons in addition to numerical indicators.

Reason 1: The large volume of the data set, which contains 16,891 articles and approximately 2 million tokens, requires detailed thematic separation.

Reason 2: In terms of interpretability, the 12-topic model is quite important in revealing the diversity of cinema studies. The 12-topic model provided a sufficient level of segmentation to reveal clearer, more consistent, and semantically distinct themes (Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity; Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema).

The selection of the number of topics was not only determined by log-perplexity and topic coherence values, but also by the semantic consistency of the study, the interpretability of the topics, and the contextual objectives of the study. Although k > 12 appears to be numerically acceptable, it was found that the topics began to consist of only a few keywords. This situation leads to excessive fragmentation in content integrity. The 12-topic model provided both statistical balance (high coherence, low perplexity) and content clarity in this context.

Table 1.

Log Perplexity and Coherence Values in the Process of Determining the Number of Topics.

Log perplexity	Topic coherence	Topic sayısı
3,049,838	0.3473	1
2,876,026	0.4090	2
2,817,287	0.4250	3
2,802,637	0.4400	4
2,800,385	0.4098	5
2,829,239	0.4445	6
2,834,329	0.4605	7
2,823,757	0.4466	8
2,826,054	0.4194	9
2,832,370	0.4429	10
2,834,990	0.4360	11
2,821,283	0.4460	12
2,831,558	0.4457	13
2,835,504	0.4288	14
2,829,442	0.4339	15
2,839,550	0.4412	16
2,826,403	0.4104	17
2,826,597	0.4282	18
2,821,117	0.4149	19
2,822,170	0.4496	20
2,823,285	0.4332	21
2,826,410	0.4369	22
2,822,974	0.4476	23
2,829,577	0.4372	24
2,835,398	0.4578	25
2,835,868	0.4347	26
2,842,354	0.4526	27
2,940,783	0.4401	28
3,299,298	0.4597	29
3,887,530	0.4274	30

Coding and Thematic Naming Process

To determine the naming of topics in topic modeling, we first examined the highly representative keywords of each theme emerging from LDA. The context of these keywords was analyzed along with the titles, abstracts, and keywords of the relevant articles. The resulting conceptual clusters were labeled by the study authors based on content similarities and academic usage. These themes were then reviewed by two expert researchers and interpreted based on their content similarities. The coding process for two sample themes is presented below. This process was supported by both qualitative content analysis and expert opinion to objectify the labeling process.

Example 1: Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity

Keywords: film, cinema, festival, african, indian, politica, india, south, censorship

Sample Article: Dovey, L. (2020). African film festivals in Africa: Curating “African Audiences” for “African Films.” Black Camera, 12(1), 13–47. https://doi.org/10.2979/blackcamera.12.1.03

Example 2: Cultural Codes and Gender Politics in Asian Cinema

Keywords: film, women, gender, cinema, chines, female, represent, cultur, queer

Sample Article: Li, J. (2017). In search of an alternative feminist cinema: Gender, crisis, and the cultural discourse of nation building in Chinese independent films. ASIANetwork Exchange: A Journal for Asian Studies in the Liberal Arts, 24(1), 86–111

Following this stage, trend analysis was conducted.

Trend Analysis

Within the scope of the study, after uncovering hidden structures and complex relationships using LDA topic modeling, the distribution of 12 identified topics over time, their volume over time (by year), their trend curves, their relative acceleration values, and the weighting ratios of topics in the large dataset were revealed using trend analysis.

In this study, trend analysis was conducted to examine the temporal trends of the themes identified through LDA (Latent Dirichlet Allocation) topic modeling. Throughout the analysis process, the volume, percentage changes, and acceleration development of the topics over time were systematically calculated in Excel. Twelve main topics identified using LDA were analyzed by dividing them into nine distinct periodic clusters (1980–1984, 1985–1989, 1990–1994, 1995–1999, 2000–2004, 2005–2009, 2010–2014, 2015–2019, and 2020–2024) between 1980 and 2024. The absolute number of articles for each topic during these periods, their percentage share in total production, and changes in acceleration over time were calculated separately.

Graphs were created to assess the temporal distribution of topics:

Volumetric Change of Topics Over Time (number of productions for each period),

Percentage Change of Topics (share in total production),

Acceleration of Topics (percentage increase rates).

Findings and Comments of the Research

Within the scope of the study, bibliometric analysis, model analysis, and trend analysis were performed in sequence. Articles indexed on cinema were identified using the WOS and Scopus databases. The term “cinema” was used as a keyword in these databases, and after the necessary screening processes were performed, a total of 16,891 articles were identified. These articles were transferred to biblioshiny, an interface of the biblimetrix package in the R programming language, and the data were classified, defined, compared, and analyzed using biblioshiny.

As shown in Table 2, 16,891 articles were accessed from 4,051 sources within the dates specified in the study, and the data set was analyzed based on this number. According to the statistical information in Table 2, 18,476 authors have contributed to scientific production related to cinema. Of these authors, 8,682 contributed individually to scientific production, while 9,794 were involved in collaborative production. Parallel to this finding, the co-authorship rate of 1.56 and the international co-authorship rate of 4.938 in the articles indicate that collaboration in articles produced in the field of cinema is relatively low.

Table 2.

Basic Statistical Data on Cinema.

Maın informatıon about data
Timespan	1980–2024
Sources (Journals, books, etc)	4,051
Documents	16,891
Annual growth rate %	12.8
Document average age	9.14
Average citations per doc	4,504
Document contents
Keywords plus (Id)	8,827
Author’s keywords (De)	37,152
Authors
Authors	18,476
Authors of single-authored docs	8,682
Co-authors per doc	1.56
International co-authorships %	4,938
Document types
Article	16,891

When looking at the growth rate related to cinema studies, it was found to be 12.8. This rate indicates that the subject is of high interest. When looking at the average age of articles published on the subject, it is seen to be 9.14. In this sense, we can say that the literature has gained momentum in recent years and that research on the subject has remained up-to-date. The annual citation rate per article is 4.504. This rate indicates the scientific value of articles published on cinema. Additionally, 18,476 authors contributing to cinema used 8,827 keywords and 37,157 keywords.

Table 3 shows the annual number of scientific publications related to cinema. In this context, it can be seen that scientific interest in the field is increasing day by day. Upon examining the graph above, it is observed that the largest numerical increase occurred in 2023, while the largest proportional increase occurred in 2007 (49.6%). This significant increase in 2007 is closely related to the integration of digital technologies into cinema. In particular, the rise of social media platforms is one of the defining dynamics of this period. Facebook’s public offering in 2006, Twitter’s launch in 2006, and YouTube’s establishment in 2005 laid the groundwork for the construction of digital culture in various forms. These developments have brought to the forefront topics such as new media narratives, user-based content production, digital aesthetics, interactive viewer experience, and online distribution strategies in cinema studies, which has led to a noticeable increase in the number of academic publications as of 2007.

Table 3.

Scientific Production Related to Cinema.

1980	6	1991	9	2002	83	2015	835	2025	1201
1981	10	1992	25	2003	113	2015	799
1982	6	1993	17	2004	140	2016	832
1983	10	1994	33	2005	187	2017	851
1984	12	1995	36	2006	244	2018	896
1985	14	1996	32	2007	365	2019	965
1986	11	1997	56	2008	330	2020	1,088
1987	14	1998	59	2009	438	2021	1,118
1988	12	1999	74	2010	516	2022	1,176
1989	22	2000	99	2011	623	2023	1,251
1990	16	2001	117	2012	723	2024	1,426*

Note. “*” was used to highlight the significant increase in the number of publications in 2024.

Among the main reasons why 2023 is the most productive year for scientific output in cinema studies are the publication of academic studies that were postponed during the pandemic, the emergence of new research topics in the field of cinema due to artificial intelligence technologies, and the transformation of content production and distribution practices on a global scale by digital publishing platforms.

Table 4 shows the annual citation values of articles related to cinema. According to this, although the average number of citations per article is relatively low, it appears to be above average between 1990 and 2006. The highest number of citations was reached in 1997 (n = 21.57), while the lowest number of citations was observed in 2024. When looking at the annual citation rates of the articles, 1997 stands out again. The highest citation rate was reached in 1997 (n = 0.74), while the lowest citation rate was observed in 1987 (0.04). When examining the citation data for 2023, the year with the highest number of articles produced, it is seen that it lags behind the years 1990 to 2006 in terms of both numerical and proportional values. When Tables 3 and 4 are examined together, it is seen that there is an increase in production in terms of numbers. However, this increase in scientific production is not receiving sufficient citations. It would not be wrong to interpret this situation as a quantitative increase that does not bring about a qualitative increase. In particular, the increasing expectations of academics regarding the number of publications by universities and the increase in the number of open access journals are the most important reasons for the quantitative increase in scientific production. This quantitative increase does not bring about methodological depth or theoretical consistency.

Table 4.

Annual Citation Number and Rates of Cinema-Related Articles.

Year	Mean total citations per article	N	Mean total citations per year	Citable years
1980	9.50	6	0.21	46
1981	9.30	10	0.21	45
1982	6.83	6	0.16	44
1983	5.20	10	0.12	43
1984	8.83	12	0.21	42
1985	4.50	14	0.11	41
1986	13.64	11	0.34	40
1987	1.71	14	0.04	39
1988	7.42	12	0.20	38
1989	4.59	22	0.12	37
1990	13.06	16	0.36	36
1991	8.67	9	0.25	35
1992	12.40	25	0.36	34
1993	10.88	17	0.33	33
1994	18.85	33	0.59	32
1995	10.78	36	0.35	31
1996	18.12	32	0.60	30
1997	21.57*	56	0.74*	29
1998	17.90	59	0.64	28
1999	17.80	74	0.66	27
2000	14.98	99	0.58	26
2001	16.90	117	0.68	25
2002	11.80	83	0.49	24
2003	15.41	113	0.67	23
2004	11.86	140	0.54	22
2005	13.63	187	0.65	21
2006	13.43	244	0.67	20
2007	7.68	365	0.40	19
2008	7.90	330	0.44	18
2009	7.80	438	0.46	17
2010	9.69	516	0.61	16
2011	7.46	623	0.50	15
2012	5.52	723	0.39	14
2013	5.00	835	0.38	13
2014	5.28	799	0.44	12
2015	4.05	832	0.37	11
2016	3.72	851	0.37	10
2017	2.92	896	0.32	9
2018	3.38	965	0.42	8
2019	3.27	1,088	0.47	7
2020	3.46	1,118	0.58	6
2021	1.75	1,176	0.35	5
2022	1.31	1,251	0.33	4
2023	0.64	1,426	0.21	3
2024	0.31	1,201	0.16	2

Note. “*” a statistical or noteworthy increase.

While universities’ focus on the number of publications in their academic performance criteria has led to a quantitative increase in scientific output, the increasing number of specialized academics in the field of cinema has also led to a quantitative increase, as well as a parallel increase in the number of academic journals. In this process, where the distinction between high-quality and low-quality publications has disappeared, it is observed that many academic journals, except for important databases, approach science in an industrial manner (Larivière et al., 2015).

According to Table 5, Alexander FEDEROV ranks first in terms of scientific contribution with 30 articles on the subject. In the above table, which lists the top 10 authors, it was found that the number of articles produced by the authors is very close to each other. The Articles Fractionalized value indicates the individual contribution percentage in multi-authored articles where authorship contributions are shared. According to this metric, it is seen that all authors are involved in joint projects. For example, Levitskaya’s fractionalized ratio is only 8.80 for 21 publications, indicating that the author is in a secondary or supporting position in most of her publications.

Table 5.

Authors Producing the Most Articles on Cinema.

Authors	Articles	Articles fractionalized
Alexander FEDEROV	30	20.75
Doru POP	27	27.00
Beti ELLERSON	23	22.33
William BROWN	22	17.87
Anastasia LEVİTSKAYA	21	8.80
Kim JEONG	20	15.70
Ewa MAZIERSKA	18	16.50
Keyan TOMASELLİ	18	14.00
Wang CHIALAN SHARON	17	11.70
Lee SANGSOON	16	9.80

According to Figure 3, Alexander FEDEROV, who has published the most on cinema, began contributing to the field in 2012 and wrote his most recent article in 2024. His most productive year was 2017. When examining Figure 3, the first publication among the 10 authors who have made the most scientific contributions related to cinema was made by Keyan TOMASELLI in 1992. Another notable point is that Beti ELLERSON is the author who published the most articles in a single year compared to other authors. ELLERSON wrote 16 of her 23 articles in 2024 alone. This indicates that ELLERSON collaborated more with other authors than others.

Figure 3.

Authors’ production over time.

Table 6 shows the average number of articles and author collaborations for the 10 countries with the highest number of publications related to cinema. According to this, the country with the highest number of publications on cinema (f = 3,255) is the United States. The United Kingdom follows the United States (f = 2,366) in terms of articles, while Germany is the country with the fewest articles among these 10 countries. The MCP (Multiple Country Publications) value in the table indicates that authors from more than one country collaborated on a publication, while the SCP (Single Country Publications) value indicates that authors from only one country were involved in a publication. The sum of a country’s SCP and MCP values indicates the total number of publications produced in that country. In this context, it is evident that the United Kingdom is the country that places the greatest emphasis on international collaboration in cinema-related articles. The countries with the lowest levels of international collaboration are India and Italy, as indicated in the graph.

Table 6.

Countries of Affiliation of Corresponding Authors and Number of Articles.

Country	Articles	Rate (%)	SCP**	MCP***
USA	3,255	19.3	3,172	83
Unıted Kıngdom	2,366	14.0	2,237	129
Spain	912	5.4	860	52
Australia	792	4.7	752	40
Canada	553	3.3	519	34
India	496	2.9	476	20
China	446	2.6	396	50
Italy	353	2.1	333	20
France	311	1.8	287	24
Germany	270	1.6	234	36

Note. “**” SCP: Single Country Publication, “***” MCP: Multiple Country Publication.

Table 7 shows the 10 most relevant institutions and the number of articles on the subject. According to this, Kings College London, University of London, University of Warwick, and University of Cambridge in the United Kingdom have made the most scientific contributions to the field with a total of 694 articles. When examining the institutions and the number of articles, it is seen that institutions in Canada also made a significant contribution after the United Kingdom. The University of Melbourne, the University of Sydney, and the University of Toronto produced a total of 407 articles. In addition to the United Kingdom and Canada, institutions in Spain, the Netherlands, and Romania were also found to be the institutions that contributed the most to the field.

Table 7.

Academic Institutions and Number of Articles on Cinema.

Affiliation	Articles
Kings Coll London	200
University of London	187
University of Cambridge	180
University of Melbourne	152
University of Complutense Madrid	134
University of Warwick	134
University of Amsterdam	133
Univversity of Sydney	130
University of Toronto	125
Babeş-Bolyai University	122

When examining Table 8, we see that “Studies in European Cinema” ranks first (f = 258) among the 10 journals that publish the most articles related to cinema. This journal, which has an impact factor of 0.2, is listed in the ESCI index. The journal “New Review of Film and Television Studies,” which ranks second with 208 articles, is a UK-based journal with an impact factor of 0.3 and is listed in the AHCI index. The journal “Studies in French Cinema” ranks third with 204 articles. This South African journal has an impact factor of 0.2 and is indexed in ESCI. The journal that publishes the most articles on cinema, but ranks last in the list, is the UK-based “Studies in European Cinema.” This journal has an impact factor of 0.3 and is indexed in ESCI. When the indexes of the journals listed in Table 9 are examined in detail, it is found that they are indexed as 9 ESCI and 1 AHCI.

Table 8.

Journals Publishing the Most Articles on Cinema.

Sources	Articles
Studies in European Cinema	258
New Review of Film and Television Studies	208
Studies in French Cinema	204
Journal of İtalian Cinema and Media Studies	202
Studies in Australasian Cinema	176
Film History	170
Journal of Scandinavian Cinema	168
Studies in Russian and Soviet Cinema	153
Black Camera	143
Studies in Eastern European Cinema	140

Table 9.

Number of Topics and Headings Revealed as a Result of Text Mining.

Topic no	Topic name	Topic terms	Rate (%)
1	Sovyet, Rus, Japon ve Türk siyah beyaz filmlerinde post kolonyal etkiler	soviet, film, cinema, black, coloni, russian, japanes, turkish, post	4.15
2	Asya sinemasında kültürel kodlar ve Toplumsal cinsiyet politikaları	film, women, gender, cinema, chines, female, represent, cultur, queer	5.50
3	Sinemada Ağ kullanımı ve Kontrol mekanizmaları	network, use, effect, control, high, trial	16.60
4	Andrei Tarkovski sinemasında sinema estetiği ve çevresel bilinç	environment, food, fire, tarkovski, modern, andrei, sur, nurs	6.10
5	Amerikan sinema endüstrisinin ulusal ve küresel kültüre etkileri	film, cinema, cultur, national, product, article, american, global, transnat, industri	2.32
6	Sinema insan ilişkisinde beden ve zaman gerçekliği	cinema, image, film, human, time, bodi, experi, cinemat, narr, reality	5.87
7	Sinemada Araştırma ve Kültürel Eğilim: Sosyal Medyanın film çalışmalarındaki kullanımı	studi, use, cinema, movie, research, social, media, cultur education	9.60
8	Sinema ve dijital medyanın kesişimi, yeni teknolojilerin film sanatıüzerindeki etkisi ve sinematik mekânın dönüşümünü	cinema, film, digit, new, media, technolog, space, video, art, image	1.60
9	Sinemada Tarihinde estetik arayışları ve film eleştirileri	film, cinema, cultur, work, narr, critic, history, aesthet	11.90
10	Savaş sonrası sinemada İtalyan ve Kore film endüstrisinde bellek ve metasinema	war, film, italian, korean, meta, cinema, memory, post, world	1.62
11	Güney Asya ve Afrika Sinemasında Politik Temsiller: Film Festivalleri, Sansür ve Kültürel Kimlik	film, cinema, festival, african, indian, politica, india, south, censorship	20.50
12	Pasolininin küresel Sinemada Etnografik ve antropolojik anlatıları	pasolini, ethnograph, anthropolog, placement, poetri, colombian, finnish, Pakistan, block	14.24

The most frequently used words in bibliometric analysis and topic modeling analysis related to cinema were also included in the study. The results obtained are presented in the word cloud map in Figure 4. According to this, the three most frequently used words other than cinema are “film, gender, documentary.”

Figure 4.

Word cloud of the hundred most frequently used keywords in cinema related articles.

Another analysis used in the study is topic modeling analysis. The topics that stand out in articles related to this analysis and the trends of these topics over time were determined using topic modeling analysis. When studies published in the field of cinema between 1980 and 2024 were analyzed using text mining methods, it was determined that the contents of these publications were concentrated in 12 main themes through the Latent Dirichlet Allocation (LDA) model.

Table 9 shows the 12 topic headings that emerged from the word-text relationship (LDA) and on which the authors agreed, along with the percentage of the data set they cover. In creating these topics, the titles, abstracts, and keywords of the articles grouped under each topic were examined in detail by the researchers.

When examining the scope of the topics, certain topics appear to be marginal. For example, while the theme “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity” has a dominant presence at 20.50%, the theme “The intersection of cinema and digital media, the impact of new technologies on film art, and the transformation of cinematic space” has a marginal presence at 1.60%. These results show that academic orientations and research interests within the field are concentrated on certain themes. In addition, themes with a marginal presence represent relatively new and developing areas such as digital media and metacinema. The low weight of such themes suggests that these areas have not yet been sufficiently explored academically or are more difficult to study methodologically. Therefore, this imbalance is directly related to the research maturity and orientations of the discipline rather than a deficiency.

According to the headings created during the topic modeling process and the topic weight ratios revealed, the topic “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity” was found to have the highest weight. In addition, the topics “The Use of Networks and Control Mechanisms in Cinema” and “Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema” also stand out from other topics with high weight ratios. The average weight of these three topics is 51.24%. As such, it can be seen that more than half of the studies related to cinema focus on a national and narrow group. Additionally, it is evident from the emerging topics that the concept of culture is also associated with cinema. These three topics, which have significantly higher weight ratios compared to other topic headings, also highlight the relationship between cinema and technology. The LDA values for these emerging topics are detailed in Figure 5 below.

Figure 5.

Topic and Keyword Distribution According to LDA Values.

The LDA word-text analysis figures are shown above. When examining the figures, the dominant keywords marked in red used in naming the 12 topics can be seen. These dominant keywords that stand out in word-text analysis demonstrate that cinema studies have a multi-layered thematic structure. This multifaceted structure reveals that cinema studies have evolved from classical formal analyses to post-structuralist, cultural studies-focused, and digital media-based approaches. Concept clusters such as “gender,”“digital,”“algorithm,”“colonial,”“festival,”“philosophy,”“body,”“covid,” and “ethnograph,” particularly visible in LDA graphs, demonstrate how closely intertwined cinema studies are with social, political, technological, and epistemological transformations.

Within the scope of the study, articles published between 1980 and 2024 were divided into nine periods and a topic modeling analysis was performed. Table 10 provides an overview of the distribution of nine different topic headings over the years. According to this, the number of articles on each topic heading has been steadily increasing. The period with the lowest number of articles is 1980 to 1984, while the period with the highest number of articles is 2020 to 2024. The topic with the highest number of articles is “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity,” while the topic with the lowest number of articles is “Memory and Metacinema in the Italian and Korean Film Industries in Post-War Cinema.” This indicates that “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity” is the most comprehensively researched topic. The topic with the least interest, “Memory and Metacinema in the Italian and Korean Film Industries in Post-War Cinema,” showed an average level of interest until 2010, but an increase in interest was observed after 2010. According to Table 10, the popularity of topics 3, 11, and 12 shows a more noticeable increase compared to other topics. The volumetric changes and trends over time in topics consisting of articles published on cinema are detailed in Figure 6 below.

Table 10.

Distribution of Subjects by Period.

table>

Figure 6.

Volumetric changes of subjects over time.

Figure 5 shows the volume change graphs of the 12 topic headings identified as a result of topic modeling. The volume change graph shows that all topic headings except for topic number 4 are trending upward in terms of volume. In particular, the 11th topic, “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity,” shows a greater increase in volume than the other topic headings. The change in trends in the topics of articles related to cinema is shown in Figure 7.

Figure 7.

Percentage and acceleration graph of topics.

Figure 7 shows the acceleration graph of the topics created within the scope of topic modeling between 1980 and 2024. According to this graph, the change trends of the articles listed under twelve topics over the years support the volume graph in Figure 6. Topics 3, 6, 9, 11, and 12 are observed to be the topics with the highest increase in publication trends. In line with the acceleration graph, “The Use of Networks and Control Mechanisms in Cinema,”“The Reality of Body and Time in the Relationship between Cinema and Humans,”“Aesthetic Pursuits and Film Criticism in Cinema History,”“Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity,” and “Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema” have seen an increase in publication trends. In contrast, the topic “Cultural Codes and Gender Politics in Asian Cinema” has followed a more stable trajectory over the years and remained at a certain level. The momentum changes in the topics covered by articles on cinema are shown in Figure 8.

Figure 8.

Accelerational change graphs of subjects.

When examining Figure 8, six of the 12 topic headings identified show positive momentum, while six show negative momentum. In particular, Topic 3, “Network usage and control mechanisms in cinema,” stands out with a positive trend. This graph also provides insight into the topics that cinema researchers will likely focus on in the future. It shows that topics with a negative trend are gradually losing popularity or that journals no longer wish to publish articles on these topics.

The significant momentum gained by the theme of “Network Usage and Control Mechanisms in Cinema” compared to other themes indicates that the impact of big data, algorithmic guidance, artificial intelligence, and digital technologies on narrative structures has increased academic interest. Another notable theme in the acceleration graph is Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema, which presents a highly specific perspective. Pasolini, who has a wide sphere of influence beyond cinema, has in recent years been frequently examined in identity and cultural studies for his literary and postcolonial critiques, his stance against Catholicism, and his ideas on sexuality and power, and this has also been reflected in cinema studies. In this context, it would be insufficient to explain the acceleration of this theme solely in terms of Pasolini. In fact, this positive momentum seems to be the result of growing academic interest in the cinematic language, cultural criticism, and anthropological narrative forms he represents.

The positive momentum of the theme “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity” is not merely an artistic trend. This situation should be evaluated as a representation of post-colonial identity searches, censorship policies implemented by authoritarian regimes, and local-global conflicts. The visibility of these geographies has increased through film festivals, cultural diplomacy, and diasporic narratives, which has brought academic interest with it.

The theme Research and Cultural Trends in Cinema: The Use of Social Media in Film Studies has seen the most dramatic decline in the acceleration graph. The most important reason for this is not only consumption, but also the natural saturation at the academic level and the emergence of new tools as important alternatives.

Furthermore, the findings show that academic interest in the theme “Cinema Aesthetics and Environmental Awareness in the Cinema of Andrei Tarkovsky” has declined. Undoubtedly, one of the reasons for this theme losing its importance in future cinema studies is the effort of a new generation of researchers to contribute to cinema studies through contemporary filmmakers.

It has been determined that the theme titled “The Effects of the American Film Industry on National and Global Culture” is experiencing a negative trend. One reason for this negative trend is that today’s filmmakers and film researchers are shifting their focus from Hollywood-centric studies to transnational, postcolonial, and local narratives. In addition, scientific studies on the effects of American cinema have now reached a point of scientific saturation after many years.

Discussion and Suggestions

The findings obtained through bibliometric analysis align with the fundamental assumptions of structuralist theory. This is because structuralist theory emphasizes intertextual relationships, recurring structures, and the reconstruction of meaning by systems. In this context, the bibliometric analysis used in this study reveals which structural patterns are repeated and which concepts come to the fore in cinema studies. In addition, the bibliometric analysis used for the study explains how structural patterns and concepts have changed over time with objective data.

Within the scope of the study, thematic orientations in cinema studies are seen to be closely related to social, technological, and cultural changes through subject modeling and trend analysis. In particular, the increase in the number of articles observed after 2010 should be interpreted not only as a numerical increase but also as an indication of cinema’s shift toward new research axes focused on cultural transformations such as digitization, identity politics, and postcolonial narratives.

Another important point that stands out as a result of the study is the subject modeling process. When the emerging themes are examined with acceleration graphs, it is seen that cinema is no longer a Western-centric art form and is represented in cinema studies in different geographical areas.

The research was conducted on two main theoretical axes. While structuralist theory focuses on the structural analysis of texts, cultural transformation theory focuses on how these structures changed over time. The findings reveal that, in line with cultural transformation theory, cinema studies have expanded in an interdisciplinary context in recent years, turning to areas such as gender, ethnography, digital technologies, and postcolonial narratives.

The study is important in terms of understanding the thematic trends in the past and future of cinema studies with the findings it presents. It also draws attention to structural transformations related to cinema and future research directions. In this context, several recommendations have been developed based on the findings of the study, and some conclusions have been drawn that could serve as a roadmap for future research.

(1) It is clear that cinema is influenced by changing cultural forms and tools. In this context, the effects of digitalization and artificial intelligence should be examined in depth at the theoretical, methodological, and thematic levels.

(2) In the context of cultural transformation, academics need to engage in interdisciplinary collaboration.

(3) The wider use of data-based methods such as bibliometric analysis in the field of cinema should be encouraged, and quality journals should give more space to such publications.

(4) Countries and authors should be encouraged to engage in international collaboration to develop strong global academic networks in cinema studies.

(5) The use of both qualitative and quantitative methods by researchers in cinema studies is important for producing objective results.

Conclusion

The data used in the bibliometric analysis and subject modeling analysis conducted within the scope of the study were obtained from the WOS and Scopus databases. Only the keyword “cinema” was used to perform a general search in both databases, and 22,061 articles were found in the WOS database and 31,358 articles in the Scopus database. In addition, some limitations and pre-screening processes were applied to the keyword search. First, only published articles were included in the study. Among these articles, only those published in English were included in the study’s dataset. Furthermore, a temporal limitation was imposed, and only articles published between 1980 and 2024 were included in the study’s dataset. After all these limitations and preliminary screening, the bibliometric analysis process was initiated with 16,891 articles.

As a result of the bibliometric analysis, it was determined that 16,891 articles were published by 4,051 different journals and that among these journals, the journal “Studies in European Cinema” kept the topic of cinema on the agenda the most with 258 articles. Additionally, when examining the other ten journals that produced the most articles related to cinema, it was observed that a significant portion of these journals were European-based. In this sense, while the sensitivity shown by European-based journals toward cinema is valuable in terms of opening up the field, the lack of a widespread journal network related to cinema poses an obstacle to the field’s development.

As a result of bibliometric analysis, it was determined that 18,476 authors contributed to cinema studies. Of these authors, 8,682 were individual authors, while 9,794 were co-authors. In this context, the co-authorship rate in articles was found to be 1.56, while the international co-authorship rate was 4.938. This result shows that authors writing articles related to cinema prefer to work individually. In addition, the low percentage of articles co-authored with authors from other countries increases the percentage of individual contributions, while also revealing an important scientific problem, namely the inability of interdisciplinary scientific cooperation to develop in the field of cinema and the inability of different ideas to come together.

The cinema theme has a continuously increasing potential with a 12.8% annual growth rate in terms of scientific production. This annual growth rate indicates that scientific interest in cinema is increasing day by day. However, despite the continuous growth in volume and academic interest in cinema, there is also the problem of the limited opportunity for articles produced to be published in high-impact journals. This is because when the impact factors and indexes of the 10 journals that publish the most articles related to cinema are examined, it is found that they are journals indexed by ESCI and AHCI, which have low impact factors.

When looking at the citation rates of articles published on cinema, it is seen that the articles have low citation rates. The average citation rate of 16,891 articles is 4.504. This situation shows that qualitative increase does not bring qualitative increase with it. The reasons for the lack of parallel development between qualitative and quantitative increases can be attributed primarily to the low level of international collaboration in the articles and the insufficient number of indexed journals such as SSCI with high citation rates in the field of cinema.

In the study, 12 topics were obtained as a result of topic modeling. In the creation of these topics, the titles, abstracts, keywords of the articles collected under each topic and the results of LDA analysis, which examines the word-text relationship, were used.

When looking at the weighted averages of these topics within the total data, it was found that more articles were written on political, cultural, and sociological representations in Asian and African cinema than on other topics. Topic 2, “Cultural Codes and Gender Politics in Asian Cinema,” and Topic 11, “Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity” are the topic headings with the highest percentage of articles written within the general data set at 26%, while Topic 8, “The Intersection of Cinema and Digital Media, the Impact of New Technologies on Film Art, and the Transformation of Cinematic Space,” is the topic heading with the lowest percentage of articles written at 1.60%.

In the trend analysis conducted as part of the study, Topic 3: Network Usage and Control Mechanisms in Cinema, Topic 12: Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema, Topic 11: Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity, Topic 9: Aesthetic Pursuits and Film Criticism in Cinema History, and Topic 6: The Reality of Body and Time in Cinema Human Relations were found to be growing rapidly, while other topics were found to be declining in momentum.

Within the scope of the study, both bibliometric analysis and machine learning were applied to the large data set, and LDA analysis revealed the above topic models. In this context, the classification and comparison of documents related to the field of cinema from the past to the present, as well as the current scientific stance, show that scientific discussions in cinema are developing based on gender equality, aesthetics, and technology. In particular, the theme titled “The intersection of cinema and digital media, the impact of new technologies on film art, and the transformation of cinematic space” clearly demonstrates the connection with technological change. However, it has been observed that concepts such as “artificial intelligence,”“machine learning,” and “AI” are used very limitedly as key terms in the analyzed articles, and that this field is still not comprehensively addressed in cinema studies. This situation indicates that there is a serious gap in the literature in the context of the relationship between artificial intelligence and cinema. Therefore, this field offers an original and open-ended perspective for future cinema studies. In this regard, one of the scientific findings revealed in this study is that the relationship between artificial intelligence and cinema has not been conceptualized and is not included in the scientific literature.

In light of all this information, the answers to the research questions are presented below, along with the evidence.

Research Question 1: What themes does scientific production in the field of cinema focus on? (Table 10, Figure 4.)

Postcolonial Influences in Soviet, Russian, Japanese, and Turkish Black and White Films: “soviet, film, cinema, black, coloni, russian, japanes, turkish, post”

Cultural Codes and Gender Politics in Asian Cinema: “film, women, gender, cinema, chines, female, represent, cultur, queer”

The Use of Networks and Control Mechanisms in Cinema: “network, use, effect, control, high, trial”

Cinema Aesthetics and Environmental Awareness in the Cinema of Andrei Tarkovsky: “environment, food, fire, tarkovski, modern, andrei, sur, nurs”

The Effects of the American Film Industry on National and Global Culture: “film, cinema, cultur, national, product, article, american, global, transnat, industri”

The Reality of Body and Time in Human Relations in Cinema: “cinema, image, film, human, time, bodi, experi, cinemat, narr, reality”

Research and Cultural Trends in Cinema: The Use of Social Media in Film Studies: “studi, use, cinema, movie, research, social, media, cultur education”

The Intersection of Cinema and Digital Media, the Impact of New Technologies on Film Art, and the Transformation of Cinematic Space: “cinema, film, digit, new, media, technolog, space, video, art, image”

Aesthetic Pursuits and Film Criticism in Cinema History: “film, cinema, cultur, work, narr, critic, history, aesthet”

Memory and Metacinema in the Italian and Korean Film Industries in Post-War Cinema: “war, film, italian, korean, meta, cinema, memory, post, world”

Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity: “film, cinema, festival, african, indian, politica, india, south, censorship”

Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema: “pasolini, ethnograph, anthropolog, placement, poetri, colombian, finnish, Pakistan, block”

Research Question 2: How have these themes changed over the years? (Table 10, Figure 5.)

The study shows that academic production in all themes remained stable until 2010. However, two important issues stand out in cinema studies after 2010. One is the shift from a stable to an accelerating trend in the number of academic articles, and the other is a paradigmatic transformation in research themes. Digital culture, gender, cultural representations, and postcolonial readings have been the most prominent themes within these paradigmatic transformations. The fact that this thematic concentration is also consistent with cultural transformation theories is one of the most important findings of the study. Structural factors such as globalization, digitalization, cultural hybridization, new media technologies, and identity politics have transformed both the production and consumption processes of cinema and the forms of academic analysis. In particular, the thematic clusters obtained from the LDA model reflect not only frequently recurring concepts but also qualitative differences in the content, form, and representational power of cinema.

Research Question 3: What topics are cinema studies likely to focus on in the future? (Figure 7.)

The study found that cinema research is likely to focus on topics such as “The Use of Networks and Control Mechanisms in Cinema,”“Pasolini’s Ethnographic and Anthropological Narratives in Global Cinema,”“Political Representations in South Asian and African Cinema: Film Festivals, Censorship, and Cultural Identity,”“Aesthetic Pursuits and Film Criticism in Cinema History,” and “The Reality of Body and Time in Cinema Human Relations” are emerging research areas.

Footnotes

Acknowledgements

The authors declare that this research was not supported by any specific funding from any public, commercial or non-profit organization.

ORCID iDs

Türker Elitaş

Serpil Kır Elitaş

Niyazi Ayhan

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

References

Bach

Krstic

Seljan

Turulja

(2019). Text mining for big data analysis in financial sector: A literature review. Sustainability, 11(5), Article 1277. https://doi.org/10.3390/su11051277

Bazin

(1967). What is cinema? University of California Press.

Bhabha

H. K.

(1994). The location of culture. Routledge.

Bystrov

Naboka-Krell

Staszewska-Bystrova

Winker

(2023). Analysing the impact of removing infrequent words on topic quality in LDA models. arXiv, 1–21. https://doi.org/10.48550/arXiv.2311.14505

Celodar

Wang

Yuan

Ç.

Feng

Jiang

Lİ

Zhao

(2018). Latent dirichlet allocation (LDA) and topic modeling: models, applications, a survey. Multimedia Tools and Applications, 78, 15169–15211. https://doi.org/10.1007/s11042-018-6894-4

Donthu

Kumar

Mukherjee

Pandey

Lim

M. W.

(2021). How to conduct a bibliometric analysis: An overview and guidelines. Journal of Business Research, 133, 285–296. https://doi.org/10.1016/j.jbusres.2021.04.070

Furtana

(2014). An evaluation of the story titled ‘Kiraz’ in the context of structuralism theory. Mavi Atlas, 2(3), 39–52.

Hansen

M. B.

(2004). New philosophy for new media. MIT Press.

Zha

(2013). Social media competitive analysis and text mining: A case study in the pizza industry. International Journal of Information Management, 33, 464–472. https://doi.org/10.1016/j.ijinfomgt.2013.01.001

10.

Larivière

Haustein

Mongeon

(2015). The oligopoly of academic publishers in the digital era. PLOS One, 10(6), Article e0127502. https://doi.org/10.1371/journal.pone.0127502

11.

Lin

T. C.

Lin

T. J.

Tsai

C. C.

(2014). Research trends in science education from 2008 to 2012: A systematic content analysis of publications in selected journals. International Journal of Science and Education, 36(8), 1346–1372. https://doi.org/10.1080/09500693.2013.864428

12.

Lune

Berg

B. L.

(2017). Qualitative research methods for the social sciences (9th ed.). Pearson.

13.

Manovich

(2001). The language of new media. MIT Press.

14.

Michelson

(1995). What is cinema? Performing Arts Journal, 17(2), 20–29.

15.

Mulvey

(1975). Visual pleasure and narrative cinema. Screen, 16(3), 6–18. https://doi.org/10.1093/screen/16.3.6

16.

Özekici

Y. K.

Ünlüönen

(2019). An explanatory basic acculturation theory of tourism-oriented cultural transformation. Journal of Travel and Hospitality Management, 16(3), 470–492.

17.

Passas

(2024). Bibliometric analysis: The main steps. Encyclopedia, 4, 1014–1025. https://doi.org/10.3390/Encyclopedia4020065.

18.

Said

E. W.

(1978). Orientalism. Vintage.

19.

Spivak

G. C.

(1988). Can the subaltern speak? In Nelson

Grossberg

(Eds.), Marxism and the interpretation of culture (pp. 271–313). University of Illinois Press.

20.

Suri

Clarke

(2009). Advancements in research systhesis methods: From a methodologically inclusive perspective. Review of Educational Research, 79(1), 395–430. https://doi.org/10.3102/0034654308326349

21.

Ültay

Akyurt