From Tweets to Polls: Linking Text Sentiment to Public Opinion Time Series

By Brendan T. O'Connor, Ramnath Balasubramanyan, Bryan R. Routledge, Noah A. Smith. International Conference on Web and Social Media, 2010

2,148 citations recorded by the source corpus.

We connect measures of public opinion measured from polls with sentiment measured from text. We analyze several surveys on consumer conﬁdence and political opinion over the 2008 to 2009 period, and ﬁnd they correlate to sentiment word frequencies in contempora- neous Twitter messages. While our results vary across datasets, in several cases the correlations are as high as 80%, and capture important large-scale trends. The re- sults highlight the potential of text streams as a substi- tute and supplement for traditional polling. consumer conﬁdence and political opinion, and can also pre- dict future movements in the polls. We ﬁnd that temporal smoothing is a critically important issue to support a suc- cessful model.

Subfields: Artificial Intelligence, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/3fee8e69b8e2df25030cf331b28b37f4dd13086c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistical and Nonlinear Physics

Citations in the corpus

Predicting political preference of Twitter users (2013, background): “The focus of studies varied from analyzing social networks of candidates and politically active users [4], [8], to predicting results of actual elections [9], [11], [15].”
Multi-Task Learning for Spatio-Temporal Event Forecasting (2015, neutral): “We detail the three models below.”
Predicting Social Unrest Events with Hidden Markov Models Using GDELT (2017, background): “Event prediction has been explored in a variety of applications, including elections [20, 21], disease outbreaks [22], stock market movements [23, 24], social unrest event prediction [11, 12, 25–31], movie earnings [23], crime [32], and failure prediction [33].”
Political Polarization on Twitter (2011, background): “and around the world (Bennett 2003; Benkler 2006; Sunstein 2007; Farrell and Drezner 2008; Aday et al. 2010; Tumasjan et al. 2010; O’Connor et al. 2010).”
Sentiment Analysis in MOOC Discussion Forums: What does it tell us? (2014, methodology): “In previous work on text-based social media sites, summary statistics derived from similarly simple sentiment analysis are demonstrated to correlate with many objective measures of population level behaviors and opinions [20, 21].”
On Using Twitter to Monitor Political Sentiment and Predict Election Results (2011, background): “also observe leading signals in Twitter sentiment, but with respect to political opinion polls (O’Connor et al., 2010).”
Identifying and Following Expert Investors in Stock Microblogs (2011, background): “O’Connor et al. (2010) propose a sentiment detector based on Twitter data that may be used as a replacement for public opinion polls.”
Social tipping dynamics for stabilizing Earth’s climate by 2050 (2020, background): “This is shown by public opinion polls on, for example, political preferences following information flows, particularly in online social media (141).”
Sentiment Knowledge Discovery in Twitter Streaming Data (2010, background): “[19] found that surveys of consumer confidence and political opinion correlate with sentiment word frequencies in tweets, and propose text stream mining as a substitute for traditional polling.”
User-Level Race and Ethnicity Predictors from Twitter Text (2018, neutral): “Through language analysis, social media data showed the potential to compliment or in part replace traditional polling (O’Connor et al., 2010), with the caveat that its demographics is not a representative sample of the real population (Culotta, 2014).”