But it seems it would be trivially easy, for somebody who knows what they're doing in Twitter's API, to scrape out all the tweets referring to each MP, and build a timeline of affect. Like, overall volume of tweets referring to an MP, relative mood of those tweets, and how that changes over time.
You'd expect that senior politicians and more vocal politicians draw more attention, both good and bad, and that politicians that are more active on Twitter also draw more attention. All of that should be able to be controlled for. And then you'd be able to check for the effect of things like changing portfolios, movement into and out of cabinet, movement into and out of opposition and all that.
Anyway - a fun project for someone with time and who knows how to play with Twitter's API. I bet the Herald's data journalists could do it in no time at all and put up some ranking that could be interesting.
Update: Thomas Lumley's had a first cut at things:
He later notes that the negative words in reply to me seem to be people sharing my outrage at the outrage-of-the-day, while negative replies to Ms Ghahraman are rather worse. It's an interesting first cut. I also really wonder how much negative engagement is driven by bots as compared to real people. David Hood points out that prior to the last campaign period things just looked less polarising. Did NZ change, or is someone messing with us?A simple example: mean number of positive and negative words per tweet that isn't just a retweet of the person in question, based on Twitter's standard 100-tweet queries pic.twitter.com/beh6mcbDv4— Thomas Lumley (@tslumley) May 21, 2019
Group accounts by party of most liked tweets and ask proportion of those accounts liking other political accounts, you get this. Shows general lack of polarisation in NZ politics (there are preferences but across a range).— David Hood (@Thoughtfulnz) January 8, 2018
Also, youth wings like other parties more than 'parents' pic.twitter.com/KyVGWWTJL5
No comments:
Post a Comment