sentiment_paper_rework
a repo containing code used for the re-work of code / data / sampling for NL and ME's sentiment paper
Steps:
- pull relevant tweets [mentioning Bernie, Hillary] from relevant timeframe [01-01-2016 - 01-03-2016] DONE
- extract overall unique users DONE
- geo-locate unique users
- create sub-samples with target users from target time frames
- visualise the composition of the sub-samples: location, gender/ethnicity, pabloscore, n_tweets[bernie, hillary]
- produce smaller [stratified?] samples for hand-labelling