Searching Racism after George Floyd

Abstract

On May 25, 2020, George Floyd, a 46-year-old African American man, was killed by a white police officer in Minneapolis, Minnesota. His death led to U.S.-wide protest that would eventually cross borders, spanning multiple continents. The transnational diffusion of protest and movement frames is a recognized phenomenon in the study of contention. Less well documented are the consequences of such protest for political interest; absent survey data, this is difficult to measure. One alternative is to use Internet search data. This visualization uses Google Trends to document variation in search interest for “racism” across 101 countries and 32 languages during the period before and after George Floyd’s killing. In the majority of countries, there was significantly increased search interest that was sustained for at least two weeks. The findings suggest that transnational protest diffusion may lead to global increases in political interest.

On May 25, 2020, George Floyd, a 46-year-old African American man, was killed by a white police officer in Minneapolis, Minnesota. Harrowing video footage capturing his death circulated widely on social and broadcast media, prompting demonstrations across the United States (Buchanan, Bui, and Patel 2020). Protest coalesced under the banner of Black Lives Matter (BLM), a movement founded in 2013 to fight racial injustice and police brutality in the United States. But protest was not confined to the United States or the BLM movement. Floyd’s death came to stand in for ongoing racial injustice worldwide, and protests would eventually span multiple continents.

Analysis to date has focused on protest incidence and social media activity related to George Floyd and BLM (“George Floyd Is Remembered” 2020; Haddad 2020). But did the protests lead to wider interest in issues of racism? Survey evidence suggests that protest and protracted movement activity can lead to enhanced interest in, and awareness of, political issues (Branton et al. 2015; Castle et al. forthcoming). Aside from surveys, another way of measuring generalized interest in topics of political interest is to use data derived from Internet search engines (Chykina and Crabtree 2018). These data are readily accessible and provide a way to gauge interest across multiple languages and countries over time.

This data visualization uses data from Google Trends. The data were collected by iteratively searching through country keyword combinations for different translations of the search term “racism.” For this, I used a Python library, “pytrents” (Hogue and DeWilde 2020), alongside Google Sheets to translate the word “racism” into target languages. ¹ Country keyword (“racism”) combinations were taken from an initial list of 244 country-keyword pairs. Data were ultimately accessible for 118 pairs, comprising 101 countries and 32 languages. Google does not provide raw search numbers; instead, search interest data are normalized before release. The normalization procedure involves taking the number of searches for the keyword in question denominated by total searches. This is then indexed on a scale ranging from 1 to 100 and normalized relative to the location and date range specified (here, January 1, 2020, to August 18, 2020, for 101 countries). Thus, a value of 100 represents the day, for the specified country-keyword combination, on which search intensity peaked, relative to other days in the same country (for further information, see Rogers 2016).

The visualization demonstrates that Floyd’s killing led to a pronounced uptick of interest in racism, and this is true across the overwhelming majority of country-keyword pairs (Figure 1). Individual tiles show daily search interest for each country on the normalized scale, where darker colors indicate increased search intensity. To determine the effective duration of increased search interest after Floyd’s killing, I also visualize, in the upper time series, the duration of time after May 25 when (standardized) mean daily search interest exceeded overall mean search interest, as well as 1 and 2 standard deviations from the overall mean. This shows that increased search interest effectively lasted about four weeks, peaking over the period from June 1 to June 4. ²

Figure 1.

Top: Time series of mean standardized search interest for “racism” across all countries. Individual points are standardized search interest for each individual country over time. The thick line is mean standardized search interest averaged across all countries. The red shaded areas show the duration of increased interest on a standard deviation scale (i.e., when search interest across all countries exceeded 1 and 2 standard deviations of the overall mean). The top time series shares the same x-axis as the bottom plot. The red vertical line indicates the date of Floyd’s killing. Bottom: Tiled heatmap of cross-national search interest in the keyword “racism”, January 1, 2020, to August 18, 2020. Search interest is displayed in intervals of 10 on the 1–100 normalized index of search interest supplied by Google. This records daily within-country variation over the specified observation period (i.e., for any given country, increases or decreases are relative to the day on which peak search volume is observed in that country). The red vertical line indicates the date of Floyd’s killing.

Existing scholarship points to the potential for movements to diffuse across borders and achieve transnational resonance (Tarrow and McAdam 2004). This data visualization suggests that episodes of mass unrest may lead to enhanced interest in issues of injustice globally. Absent more granular search data, we cannot make inferences at the individual level. But here, several questions would be worthy of study. Does increased interest translate into sustained increases in political awareness? What are the sociodemographic correlates of those who express heightened political interest after such movement-led campaigns? And do such campaigns lead to heightened engagement among the politically uninterested as well as those who had prior interest in politics?

Supplemental Material

sa – Supplemental material for Searching Racism after George Floyd

Supplemental material, sa for Searching Racism after George Floyd by Christopher Barrie in Socius