‘It’s a fiver for BJP in Delhi’

Five seats for BJP and two for AAP in Delhi.

This is what Technopark-based big data analytics player Flytxt predicts, based on analysis of available public data using a mathematical model developed in-house.

The only other State surveyed is Kerala where it predicts a slight edge for the Left Democratic Front (LDF) with 10 seats against nine for Congress-led United Democratic Front (UDF) and one for others.

Since election outcome is an aggregated preference of voters, a possible method of predicting the result is to estimate voting preference through statistical models.

This is done by making use of historic data and changing user preferences across time expressed through opinion polls and social media.

In the US Presidential elections, Barack Obama’s campaign made heavy use of data sciences to understand people’s opinion and its temporal change closely so that campaigns could be optimised.

“Here, our attempt is to make use of similar predictive models in the Indian context to predict election outcomes in advance by using our expertise in data sciences and mathematical modelling,” says Jobin Wilson, architect at research and development at Flytxt.

This version of the model can be applied to obtain forecasts for other States also and used to estimate total seat count for the Lok Sabha.

“However, since this is our first attempt at election forecasting by leveraging our analytics expertise, Kerala and Delhi have been chosen as test cases for validating our model.”

Delhi was chosen especially because of its strong social media presence.

Tweets have been used in the analysis since Twitter is a heavily used platform for online political deliberation where people express their opinions/views about political issues.

The sentiment of a sample of random users towards political parties was obtained by analysing their tweets using standard ‘natural language processing’ techniques.

The final sentiment score for each tweet was calculated by taking into account the importance of the tweet and that of the person who does it.

Raw data on prior Lok Sabha and Assembly elections was sourced from the Election Commission Web site. Opinion poll results were from Wikipedia.

Facebook posts were not considered for purposes of the survey since they were restrictive in terms of access.

But the word limit and compact size of tweets and their structuring more than served our purpose, Wilson said.

COMMENT NOW