It is proving extraordinarily difficult to run Sentiment Analysis successfully for such a complex UK General Election.
However, some trends we have picked up:
– the Liberal Democrat vote is much steadier (though still generally in decline) in constituencies with a well-known candidate (not always necessarily the incumbent), giving them probably 30-40 seats;
– the Conservatives are almost bound to gain at least one seat in Scotland almost by chance due to the split against them (although they may lose the one they hold);
– the DUP vote will be down on 2010, although unevenly depending on context (perhaps more so in safer seats; thus probably not enough to cost them any, although it is tight);
– UKIP will probably win 3-4 seats (Farage, Cardwell, perhaps Reckless plus probably one more by chance on an even three-way split against them); and
– the Greens are too hard to pick up geographically (it is hard to run sentiment with precision as their target seats are usually in divisions of a city – e.g. a particular Brighton seat or a particular Norwich seat), but there is no reason to doubt their polling score of about 5-6% in England.