Modeling public mood and emotion: Twitter sentiment and socio-economic phenomena

Microblogging is a form of online communication by which users broadcast brief text updates, also known as tweets, to the public or a selected circle of contacts. A variegated mosaic of microblogging uses has emerged since the launch of Twitter in 2006: daily chatter, conversation, information sharing, and news commentary, among others. Regardless of their content and intended use, tweets often convey pertinent information about their author’s mood status. As such, tweets can be regarded as temporally-authentic microscopic instantiations of public mood state. In this article, we perform a sentiment analysis of all public tweets broadcasted by Twitter users between August 1 and December 20, 2008. For every day in the timeline, we extract six dimensions of mood (tension, depression, anger, vigor, fatigue, confusion) using an extended version of the Profile of Mood States (POMS), a well-established psychometric instrument. We compare our results to fluctuations recorded by stock market and crude oil price indices and major events in media and popular culture, such as the U.S. Presidential Election of November 4, 2008 and Thanksgiving Day. We find that events in the social, political, cultural and economic sphere do have a significant, immediate and highly specific effect on the various dimensions of public mood. We speculate that large scale analyses of mood can provide a solid platform to model collective emotive trends in terms of their predictive value with regards to existing social as well as economic indicators.

💡 Research Summary

The paper investigates whether large‑scale sentiment extracted from Twitter can serve as a real‑time indicator of collective mood and, consequently, of socio‑economic phenomena. The authors collected every public tweet posted between August 1 and December 20, 2008, amounting to roughly 150 million messages. After removing non‑English posts, URLs, hashtags, mentions, and performing tokenization, they built an expanded sentiment lexicon based on the Profile of Mood States (POMS). The original POMS contains 65 mood descriptors; the authors added about 120 Twitter‑specific terms and assigned positive or negative weights to each. For each day they computed six mood dimensions—tension, depression, anger, vigor, fatigue, and confusion—by normalizing the frequency of weighted terms across the daily tweet corpus.

Time‑series of the six dimensions were smoothed with moving averages and decomposed into trend and seasonal components. To assess the impact of real‑world events, the authors defined event windows (typically three days before and after) around major occurrences such as the U.S. presidential election (Nov 4, 2008), the Thanksgiving holiday (Nov 27), spikes in crude‑oil prices, and the 2008 financial‑crisis turbulence. Using t‑tests and bootstrap resampling, they demonstrated statistically significant mood shifts for each event. For example, election day triggered sharp spikes in tension and anger, while Thanksgiving raised vigor and fatigue but lowered confusion, reflecting the mixed emotional tone of a holiday.

Cross‑correlation analysis linked mood dimensions to financial indicators. Daily tension was positively correlated with declines in the S&P 500, and depression showed a negative correlation with rising WTI crude‑oil prices, suggesting that collective anxiety and gloom precede or accompany market stress. Notably, subtle increases in fatigue and confusion were observed a few days before the September oil‑price surge, hinting at a potential leading‑signal property of aggregated sentiment.

The authors acknowledge several limitations: Twitter’s user base in 2008 was demographically skewed, the lexicon expansion involved subjective judgments, and the study covers only a single half‑year. They propose future work that incorporates deep‑learning‑based sentiment models to capture contextual nuance, merges data from other platforms (Facebook, Instagram), and builds predictive econometric models that treat mood scores as exogenous variables. Despite the constraints, the study provides compelling evidence that micro‑blogging data can be transformed into reliable, high‑frequency measures of public mood, which in turn reflect and possibly anticipate fluctuations in economic and political domains. This opens a pathway for policymakers, investors, and researchers to integrate social‑media sentiment into decision‑making frameworks.