Introduction

        Social media such as Twitter and Facebook is now days a powerful communication platforms for idea sharing, breaking news, personal networking, political opinions and following someone. By using smart phones people can communicate and coordinate their activities geospatially in real time. Twitter is a popular online micro blogging service established in 2006. User can write and post short messages up to 140 characters of maximum size for each post to their followers in Twitter, these short messages called tweets, which can be searched by keywords, author name and hashtag (Tsou and Yang 2012), we will use these keywords as our primary data for searching (Nunes, Lopes, and Abdelouahab 2012). Have used the an approach based on Twitter opinion analysis about politics in order to collect information and determine what is the opinion of people about current election.

       With the increased awareness about the capabilities of Geographical Information Systems(GIS), it has been used for newer and newer application areas; including elections, a recent example is of US elections where GIS is used to show geographic count of votes and candidate support. GIS allows storing database records that are geo-referenced e.g. latitude/longitude information linked to the data (Nunes, Lopes, and Abdelouahab 2012). Among open source solutions to deal with spatial data is the QuantumGIS (QGIS). QGIS provides a number of plugins developed by the community supportive of open source software’s, among these plugin is a plugin named twitter2QGIS (currently in experimental condition). It requires a twitter account, twitter application account, Tweepy API package which will be installed first. This plugin allows the user to specify the keywords for searching the tweets and returns the results in form of the shapefile which fulfills the most common barrier of data unavailability in case of most GIS projects.

       Once the data is available, it can be subjected to multiple types of spatial analysis as per the requirement at hand e.g. to show the intensity of spatial distribution for point data i.e. geo-located tweets in this case, heat maps can be generated. Heat maps show the results in the form of hot spot (dense cluster), cold spots (sparse cluster) and regions in between, this helps the analyst to quickly analyze the regions having more number of geo-located tweets and similarly regions of less number of tweets. Interestingly these results can be used to analyze the public support for different political parties and the heat map results when spatially joined with the electoral boundaries can help to identify the areas e.g. union councils where more and intense campaigns are required to increase the party popularity and support and on the other hand efficiently spend or save time and cost for areas which already show relatively greater support.