Incorporating mobile application usage data into electoral models significantly improves the ability to predict municipal voting patterns compared to relying on socioeconomic factors alone, according to an analysis of 4,102 urban and suburban localities in France during the 2019 and 2024 European elections. Led by researchers at IMDEA Networks in collaboration with Universidad Carlos III de Madrid, Northeastern University, and Orange Research, the study demonstrates that digital content consumption habits are increasingly correlated with electoral outcomes across French municipalities.
The investigation evaluated two comparative statistical models for each political party contesting the elections. The baseline model relied exclusively on established socioeconomic indicators, including local income levels, unemployment figures, and age demographics across the analyzed municipalities. The expanded model integrated mobile network data reflecting application usage alongside those traditional demographic variables, allowing researchers to evaluate the precise predictive contribution made by mobile digital activity.
Adding mobile application usage data produced a notable enhancement in model performance across both election cycles. The inclusion of digital metrics improved the correlation between the vote shares estimated by the models and actual electoral results by approximately 10 percent in the 2019 election and by roughly 28 percent in 2024, calculated as an average across all political parties. This marked increase indicated that mobile traffic captures distinct spatial patterns in voting behavior across municipalities that traditional socioeconomic indicators alone fail to detect.
The researchers observed that while traditional socioeconomic factors maintained a steady importance across both European elections, the statistical relationship between digital consumption habits and election outcomes intensified substantially over the five-year period. Marco Fiore, a research professor at IMDEA Networks, highlighted this divergence as an important quantitative signal that social media and digital content consumption may be playing an increasingly significant role in the ongoing dynamics of political voting and polarization.
Detailed analysis of specific digital platforms uncovered distinct partisan alignments across the analyzed French localities. Localities with elevated Facebook usage exhibited higher vote shares for the Rassemblement National, the French far-right party, across both the 2019 and 2024 European elections. Concurrently, above-average Facebook activity in a municipality was associated with lower vote percentages for competing parties, a pattern that proved particularly pronounced for the left-wing La France Insoumise in the 2024 contest.
Usage of the platform X, formerly known as Twitter, displayed a noticeable temporal shift between the two electoral cycles following its change in corporate ownership. In 2019, higher relative usage of the platform was concentrated in localities where Emmanuel Macron’s coalition, Renaissance, and the Green party, Europe Écologie, achieved stronger electoral results. By 2024, however, greater local engagement on X correlated with higher vote shares for La France Insoumise, alongside reduced support for the Rassemblement National and La France fière.
Instagram demonstrated an alternative geographic and political pattern distinct from that of Facebook. Municipalities with higher Instagram consumption relative to national averages corresponded to areas where moderate and centrist political parties performed more effectively. In contrast, localities recording stronger support for the Rassemblement National exhibited systematically lower levels of Instagram activity.
The five-year window between the two elections coincided with major transformations in the broader mobile app ecosystem. Orlando E. Martínez-Durive, a senior research engineer at Net AI who conducted the work as a postdoctoral researcher at IMDEA Networks, noted that TikTok was not among the mobile services measured in 2019, yet by 2024 it generated more network traffic than any other application examined. Meanwhile, existing platforms such as Instagram, Facebook, and YouTube adapted their offerings by deploying short-format video features including Reels and Shorts, reflecting rapid shifts in consumer phone habits.
Methodologically, the research team analyzed the volume of network traffic generated by each mobile service across individual localities and compared these figures against national averages. The authors emphasized that all mobile data analyzed were fully aggregated by mobile network operators to preserve individual privacy, making it impossible to determine how any individual citizen cast their ballot. Because telecommunications operators continuously measure aggregate traffic to manage infrastructure, the approach provides an independent view of societal digital behavior that does not rely on proprietary data releases by tech platforms.
Building upon earlier work from 2021 showing that mobile traffic mirrors local socioeconomic profiles such as income and education, assistant professor Iñaki Úcar of Universidad Carlos III de Madrid affirmed that app usage provides an additional layer of insight into voting patterns that standard demographic metrics omit. The study authors noted that this privacy-preserving framework could be adopted by public authorities and regulatory bodies across other countries to track digital habits around election periods and gather empirical evidence regarding the relationship between digital platforms and political polarization.
The cartographic and spatial comparisons presented by the research team further illustrate the geographic alignment between digital service consumption and partisan voting behavior throughout France. In their geographic mapping of the electoral data, researchers evaluated the spatial distribution of votes cast for the Rassemblement National alongside municipal variations in Facebook activity. On these comparative maps, darker geographical regions represent localities with higher vote shares for the far-right political party. Correspondingly, the Facebook usage distribution uses color-coded metrics, where red shading denotes digital application activity exceeding the national average, and blue shading highlights localities where mobile traffic falls below the nationwide benchmark. Together, these two spatial distributions reveal a clear geographical association: municipalities where Facebook usage exceeds the national average tend to record stronger electoral support for the far-right party.
This passive analytical approach provides a valuable methodological complement to conventional opinion surveys and demographic polls. Aggregate telecommunications telemetry, by contrast, records the digital applications and services that populations actively use in everyday life and tracks how those collective usage habits transform over time. Because mobile network infrastructure measures data transmissions for each application using an identical technical framework, the researchers were able to monitor shifting platform engagement across thousands of individual localities from one European election to the next. This consistent measurement across time and geography enabled the research team to conduct fine-grained comparisons between evolving mobile traffic habits and actual municipal voting results at an unprecedented national scale.
The broader implications of these findings extend well beyond the French electorate, carrying particular significance for impending democratic processes across Europe, including Spain’s approaching general election. As political and social polarization becomes increasingly noticeable throughout modern democracies, analyzing how citizens gather information and what media formats they consume provides indispensable insight into the shifting dynamics of partisan alignment. The collaborative study by IMDEA Networks, Universidad Carlos III de Madrid, Northeastern University, and Orange Research indicates that political preferences are increasingly mirrored in broad digital content consumption patterns rather than restricted solely to direct political discourse. By assessing how much network traffic individual applications generate across localities relative to national averages, researchers can trace structural changes in civic alignment, observing how shifts in digital communication ecosystems coincide with electoral realignments across urban and suburban populations.
Subject of Research: Correlation between mobile application data consumption and municipal voting patterns in French European elections
Article Title: A new study reveals the growing link between digital consumption and political preferences
Article References: A new study reveals the growing link between digital consumption and political preferences. Original publication (Bibliographic details unavailable.)
Image Credits: AI Generated
DOI: Not provided
Keywords: IMDEA Networks, European elections, mobile data, digital consumption, political polarization, socioeconomic indicators, electoral modeling, Facebook, Rassemblement National, X platform, La France Insoumise, Instagram, TikTok, aggregated network traffic, French politics, voting patterns
Cite Scienmag News
Courtney Benton. (November 18, 2026). Mobile App Usage Data Increasingly Correlates With French Municipal Voting Patterns. Scienmag. https://scienmag.com/mobile-app-usage-data-increasingly-correlates-with-french-municipal-voting-patterns/
Courtney Benton. "Mobile App Usage Data Increasingly Correlates With French Municipal Voting Patterns." Scienmag, 18 November 2026, https://scienmag.com/mobile-app-usage-data-increasingly-correlates-with-french-municipal-voting-patterns/. Accessed 18 November 2026.
Courtney Benton. "Mobile App Usage Data Increasingly Correlates With French Municipal Voting Patterns." Scienmag. November 18, 2026. https://scienmag.com/mobile-app-usage-data-increasingly-correlates-with-french-municipal-voting-patterns/








