Splunk the Vote: BBC Election Debate

This post is the first in a series analysing social data about the UK General Election 2015.

BBC leaders debate

The third official debate has come and gone – this time without Cameron and Clegg. Perhaps this is why we saw the fewest tweets (179,000) collected during the debate compared to the previous two debate (216,000 & 312,000).

But how did the two leaders compare to those in the five opposition leaders who took part in the debate?

In the third part of my #SplunkTheVote series I took to Splunk to find out.


Cameron Tweets

4 key findings:

  • Twitter was still talking about Cameron. We collected 19,000 tweets mentioning his name despite his absence.
  • There was a big spike in negative sentiment for Farage when he branded the audience as “left-wing”.
  • Miliband had the highest overall sentiment from Twitter users with only 8% of tweets being negative.
  • Clegg faired paticulalary badly during this debate with the most negative tweets posted (26%).

The numbers

  • Ed Miliband (Labour)
    • Tweets: 25,596
    • Overall Sentiment:Positive (92%)
  • Nigel Farage (UKIP)
    • Tweets: 25,839
    • Overall Sentiment: Positive (86%)
  • Nicola Sturgeon (Scottish National Party)
    • Tweets: 22,380
    • Overall Sentiment: Positive (88%)
  • Leanne Wood (Plaid Cymru)
    • Tweets: 4,446
    • Overall Sentiment: Positive (81%)
  • Natalie Bennett (Green)
    • Tweets: 8,723
    • Overall Sentiment: Positive (87%)
  • David Cameron (Conservatives) – absent from debate
    • Tweets: 18,781
    • Overall Sentiment: Positive (88%)
  • Nick Clegg (Liberal Democrats) – absent from debate
    • Tweets: 5,221
    • Overall Sentiment:Positive (74%)

The data

If you want to build some cool stuff with this data, grab the data set here and let me know what you find.

I extracted this data from my Splunk instance with the following API request:

>curl -k -u admin:changeme --data-urlencode search="search index="ge2015" source="rest://twitter_hashtags" #bbcdebate earliest="04/16/2015:20:00:00" latest="04/16/2015:21:30:00" | fields _raw" -d "output_mode=csv" > 1604leadersdebatetweets.csv