Indian Elections 2014 – Big Data for Billion People!
India recently concluded their general elections. What comes to our minds as Indian diaspora when we think about Indian Elections? Humungous rallies, colorful campaigning, myriad party symbols and boisterous victory laps. The Indian General Elections also have another perspective which often does not figure in our most buoyant thoughts. - 300 parties, 8000 candidates, 800 Million voters, 1 Million booths served/secured by ~20 Million officials. The heady mix is further embellished with variety of structured & unstructured information - candidate histories, crime records, declared assets and audacious election manifestos. Mixed with the above is the frenetic activity on the day of results. Live streaming of results: ~21000 votes to be counted per second, from all corner of the country spanning an area of ~1 million square miles. Thus, elections in India is a classic Big Data problem and the 2014 general elections was the biggest of them all.While technology may be able to process this humongous data, how can all this information be consumed and understood by a billion people? That too, in real time as it happens? Gramener took the task of analyzing such large volumes of data into meaningful information for the general public to consume easily. Following is an account of how this was achieved. (www.gramener.com) The election timeline had 3 logical phases spread over 2 months: 2 weeks of Pre-elections, Polling of 5 weeks and Post-election phase of 2 weeks. Even before thinking about the imminent elections was the task of assimilating historical data. Data from 1950s was to be gathered from multiple PDF files published on the Election Commission of India (ECI) website and followed with the task of cleaning, correction & collation. Synthesizing this data with various other credible data sources completed the painstaking exercise of building an integrated, structured elections data source. This master data source served as the backdrop for 2014 election with over 60 years of election data. Gramener’s proprietary analytics and visualization technology was used to complete this task and this history was hosted through a web page (https://gramener.com/election/parliament). This first step brought about an interesting twist to election reporting. Ability for people to consume data stories of over 60 years. Print, Online and On-Air media was agog discussing these stories since audience was able to understand simple but powerful information easily – victories with highest margins, regions who always unseated the incumbent, most persistent yet unsuccessful parties, highest number of candidates contesting for a single seat, counter-intuitive stories like “victory margins of the winner goes up when number of candidates go up” and so on. For the first time, the electorate in India was exposed to data-based reporting with hard facts and numbers, which were easy to consume as data stories. Then the polling phase began. The same analytics and visualization techniques were used to report the 5 weeks of polling, by providing live facts on voter turnouts, exit poll surveys etc. Thus, the stage was set for the grand finale – the counting day. Torrential stream of vote-by-vote count and results were to seal the fate of candidates and parties. Live data feeds need to be analyzed, visualized and data stories need to be reported real time on national television! Accuracy, speed and ease of consumption will be put to test in front of the entire nation which had reached a crescendo in anticipation. Gramener’s analytics and visualization engine, Gramex® was programmed to arrive at an election results dashboard which were to meet these heavy expectations. Particular care was taken to enable real-time slice/dice of data for deeper insights as it streamed in. It has to be novel and farthest from plain reporting of live numbers which had been the norm until then. With plans in place and trial runs completed, the visualization dashboard went live on the morning of 16th May - the counting day. How will the people of India receive this? Will they like it? Will they understand it? Gramener and CNN were holding their breaths to know the reception of this by India, laced with the trepidation - will Gramener technology stand the ultimate performance test on this D-Day? Answer came in the first hour – the visualization dashboard had over a million hits in an hour. Grew to a total of 10 million hits in a span of 12 hours! Social media exploded with humongous response to these live visualizations! Common people started consuming and sharing deeper insights real time! The various mass media channels lapped up the insights and reported them across the county. A new era in Indian election reporting was born! (https://ibn.gramener.com/live). It was not only a new era in election reporting, but it was an epitome of collaboration between mass media and technology companies. Charged with this confidence, Gramener now is getting ready for another large democracy preparing to use its right to select future leaders – US elections!
