Context

Canadians elect representatives to the House of Commons. The leader of the party who has the confidence of a majority of members of the House forms the Government. This data explores the results of elections over the past couple of decades.

Content

This dataset includes the election results from seven elections over the past twenty years, and has been updated to reflect the 2019 election. Election results are provided at the polling-station level. As Canada is a bi-lingual nation, some variables are given in both English and French. Some of the variable names have been left overly long and will want renaming (let's face it, it wouldn't be a data science project if there wasn't at least some data cleaning involved).

I have identified errors in the 1997 raw data file that I downloaded from the Elections Canada website. These errors involve incorrect tallies of votes by poll, such that the candidate riding totals do not match the actual candidate votes by riding for one riding that has been identified (and possibly others, as yet unidentified). Consequently, the results file does not currently include results from 1997, however information regarding this election remains in other files, as the identified errors do not involve the election date, riding names, or candidate details, for example. If a correct version of the poll-by-poll results data is located for the 1997 election, I will update the results file and re-upload.

Acknowledgements

All data contained within this dataset was taken from the Elections Canada website, with some wrangling to ensure consistency across elections. The data is a reproduction of data available from this link: https://elections.ca/content.aspx?section=ele&dir=pas&document=index&lang=e

Licensing information is provided at this link: https://www.elections.ca/content.aspx?section=pri&document=index&lang=e

Inspiration

Elections data has many applications, looking at party support over time, or the geographic trends of party support within a single election. The incumbency indicator may be useful for exploring the extent to which incumbency may benefit a candidate. Candidate occupations may lend themselves to some sort of text analysis or wordclouds.

Future Intention

I aim to update it with information regarding returning officers. I also hope to update it with by-election information.

Additional Information

Note that in the Ridings.CSV file, there are two sets of district names. The first use two hyphens to separate place names, while the second uses an em dash. The first will be more useful for filtering, etc, in your Kernels, while the second is better for outputs and reports.