Dataset of Jaccard similarity indices from 1,597 European political manifestos across 27 countries (1945-2017)

William Sanger; Thierry Warin

doi:10.1016/j.dib.2019.103907

Dataset of Jaccard similarity indices from 1,597 European political manifestos across 27 countries (1945-2017)

Data Brief. 2019 Apr 9:24:103907. doi: 10.1016/j.dib.2019.103907. eCollection 2019 Jun.

Authors

William Sanger¹, Thierry Warin²

Affiliations

¹ Polytechnique Montréal, CIRANO, Canada.
² SKEMA Business School, CIRANO, Canada.

Abstract

This dataset compiles the results of our computations of Jaccard similarity indices of political manifestos from 27 European countries since 1945. The raw text is accessible through the Manifesto Project's API in R. In total, 1597 political manifestos have been compared on a country level, providing measures of similarity between different political platforms. In addition to the Jaccard similarity indices gathered into 27 csv files (one per country), the R code to download, transform, and compare political manifestos is provided. Moreover, we also provide the R code to execute all the necessary computations for a whole country.

Keywords: Data science; European elections; NLP.