And the LiveDebate winner is…

The LiveDebate on Tuesday certainly gave the nation food for thought, and gave the Text Mining Solutions algorithm a lot of new data to process. Over 30,000 people tweeted over the two days following the debate, all sharing opinions; likes and dislikes, approval or disapproval and happiness or rage, regarding what each party leader has said. TMS analysed all of them, and can now provide you with an answer that the polls cannot. In order for the polls to declare a winner of the debate, as it would require new questions of the electorate. This is an issue that TMS does not have, as the database already has the new opinions on the new topic.

By filtering the (now 21M!) processed tweets to posts mentioning only Labour or Conservatives, and the hashtags #itvdebate or #leadersdebate, a specific and focussed pool of data was selected. This meant that conversations surrounding the debate (such as the LibDems missing invitation), that were not about the hour-long programme and what occurred specifically, were not included.

From this refined group, the algorithm is then able to extract the sentiment, both the polarity and strength, of the tweets. This allows a comparison between the Conservative opinion and the Labour opinion, based solely on the leaders’ performance in the debate. Following this analysis, every tweet was dropped into a graph, which is the one you see below.

Text Mining Solutions believes it shows a Labour win!

Every dot represents a single tweet, and its size demonstrates how many times it was retweeted, and therefore its popularity. The red dots represent tweets about Labour, and the blue are Conservative. The x-axis (across), displays positive sentiment which runs from 1 to 5, left to right. So, any dots the the far right of the chart are highly supportive. The y-axis (vertical) shows negative sentiment, running from – 1 to -5  so the top of the chart is highly critical of one party or another. We have removed the dots from the lower left hand corner for clarity since these displayed neutral sentiment, i.e. where tweeters have made comment but have not expressed an opinion either way.

At first glance, the distribution is fairly even. However looking closely, it appears that there are a greater number of red dots toward the right of the chart than blue, and a greater number of blue dots toward the top of the chart than red. Therefore TMS is able to conclude that Jeremy Corbyn won the debate, due to the greater amount of tweets conveying positive sentiment towards the Labour party.

Processing and analysing tweets following Tuesday’s LiveDebate has been a brilliant opportunity for Text Mining Solutions to demonstrate the level of detail that the algorithm is able to attain, down to text sentiment. It has also provided the public with an answer which the polls are unable to provide.