Predicting 2016 US presidential election polls with online and media variables

Veikko Isotalo*, Petteri Saari, Maria Paasivaara, Anton Steineker, Peter A. Gloor

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterScientificpeer-review

7 Citations (Scopus)


Traditional media has always played a large role in elections by informing voters and shaping opinions, and recently, social media and various Internet information sources have also become considerable influencers on the voters. There is data publicly available on how these information sources and media channels are being used, which could potentially be analyzed for their effects on the election process. This chapter aims to determine if social media, Internet traffic, and traditional media data can be used to predict elections by searching for patterns between the data and poll numbers for 2016 US Republican and Democratic primaries. The results suggest that machine learning models with linear regression can produce quite accurate predictions; also statistically significant correlations were found between polls and betting odds and polls and Facebook page likes. More sophisticated methods could allow for better forecasting using this publicly available data.

Original languageEnglish
Title of host publicationDesigning Networks for Innovation and Improvisation
Subtitle of host publicationProceedings of the 6th International COINs Conference
Number of pages9
ISBN (Electronic)978-3-319-42697-6
Publication statusPublished - 1 Jan 2016
MoE publication typeA3 Part of a book or another research book

Publication series

NameSpringer Proceedings in Complexity
ISSN (Print)2213-8684
ISSN (Electronic)2213-8692


Dive into the research topics of 'Predicting 2016 US presidential election polls with online and media variables'. Together they form a unique fingerprint.

Cite this