Analysing aggregated UK polling data
- The tweet collection & storage pipeline used in email-service is used to also collect tweets from @britainelects.
- extract_britain_elects.py queries the necessary data from my MongoDB database and saves it. Projections allow for querying specific fields within each document, keeping only the tweet body and timestamp in this case.
- The additional polling historical data in uk_polling_report_historical.csv was scraped from UK Polling Report.
- polling_report_history.py contains a function to clean this data to combine with the Westminster Voting Intention Twitter data from Britain Elects.
- analyse_polling_data.ipynb contains various bits of analysis using the combined data, as well as a quick look at the sort of text found in general within all of Britain Elects' tweets, including those unrelated to Westminster Voting Intention. This data can be found in the britain_elects_all folder.
