Electoral Intelligence

THREE SCREENS BLINK all day long at the Gurugram- based analytics, strategy and political consulting firm, N-Sight Consulting. Five techies rev up the machines and 30 full-time number crunchers feed in live data as one of the boxes comes alive with a vivid, real-time picture of Mumbai South where a high liberal sentiment seems to pip the overall conservatism. Karan Singh, 41, the founder and chief executive of N-Sight, is at the centre of the digital orchestra. He claims to have worked for “hundreds of politicians” with 89 per cent accuracy over the last seven years of polling pan-India—from the civic level up to the ongoing General Election. Big data slices and dices constituencies to reflect demographic truths that the politician-client can act upon with tailor-made messaging. Other screens blink on with static secondary and primary data. The firm deploys some 3,000 contract workers to collate data on the ground over tablets, which are being constantly fed into the firm’s ‘engine’ server in what the Oxford and London School of Economics-educated Singh dubs the “war room”.

Singh is part of the extraordinary role that technology has begun to play in Indian elections. If in 2014 it was 3D holograms, this time big data and Artificial Intelligence are being used by parties and candidates to micro-target voters with powerful and customised messages. The electorate’s tastes and preferences are trapped from, say, digital wallets or utility bills. Voter data is then organised by computer algorithms for campaigns to quickly sift and sort through issues voters are passionate about. Predictive modelling is the flavour of this season. Simultaneously, social media has evolved into an all-consuming platform since the 2014 Lok Sabha elections. The country has 450 million smartphone users, of whom 200 million are on WhatsApp and about half a million on Facebook, making social media the primary tool for politicians and parties to act as influencers in the polls. It is becoming increasingly impossible to conceive of a strong campaign without analytics, big data, AI and a social media strategy. Smartphones and apps enhance the technology and a stream of apps, like Neta App or Next Election, can be accessed from handsets.

While consumption trends and behavioural patterns are being tapped to enrich datasets, availability of data from the interiors of the country is still a stretch. But Prodyut Bora, the man who started the Bharatiya Janata Party’s (BJP) IT cell in 2007, is upbeat. In this election it might still be nascent but he believes a new range of technologies could be game changers in 2024. “In the video feed of a political leader, the programme code will figure out the colour you like and depending on your preference, the kurta colour of the politician changes,” says Bora.

Bora quit the BJP in 2015 and floated his Liberal Democratic Front in Assam two years later. But he was the national convener of the BJP’s IT Cell when the party launched 3D hologram rallies in 2014. The technology was then used to hold 1,350 3D rallies at the state and constituency levels. Why don’t we see such ‘Beam me up, Scotty’ innovations this time around? “In 2014, holograms were a novelty. Today, with 4G, people are talking over video phone all the time and there is no novelty in that, just like a 4K high- definition video is no big deal,” he says.

ELONNAI HICKOCK, CHIEF operating officer of the Centre for Internet and Society that has offices in Bengaluru and Delhi, echoes the sentiment differently in a paper titled ‘The Influence Industry: Digital Platforms, Technologies, and Data in the General Elections in India’: ‘Though other elections have leveraged digital platforms prior to 2014, the 2014 general elections in India marked a change in electioneering strategy and have been said to be different than previous electoral campaigns in India. Traditionally, candidates have relied on local and cultivated knowledge and have used techniques such as opinion polls, voter banks, door to door campaigning, booth wise caste and religious profiling, key community leaders, and volunteers to focus on groups of voters. In contrast, in the 2014 elections, the use of data, technology, and digital platforms played a central role in the way the campaigns were designed, structured, targeted, implemented, and communicated. Technology and data were also key in strategically navigating the complexities of demographics, religion, politics, caste, geography and community in India.’

Most such technologies originated prior to the US presidential elections. Ankit Lal, social media and IT strategist for the Aam Admi Party (AAP), says this is what happened with social media in 2015-16 before the US’ last election. He expects the next wave somewhere in the middle of the year which will then creep into countries like India. For now, Lal has his hopes on a few of them. Virtual Reality (VR) is one such variant which the party tested in March. “On the app, without any human interface whatsoever, you can take a selfie with [party chief] Arvind [Kejriwal] over the phone from any location including a personalised message from him,” says Lal, who used UK-based VR solutions provider Zappar to test it.

He says 3D holograms, which they experimented with, didn’t take off because costs were prohibitive, even if mobile 3D vans were used to go live in various locations. “It costs Rs 15 lakh to set up such a vehicle,” he says. In 2014, the BJP in particular took a leaf out of Barack Obama’s US presidential campaign in 2009—that stands out for its use of predictive analysis powered by big data to model outcome scenarios based on early voting data. Start-ups like Voxta, a speech recognition service which helped in relaying of Modi’s speeches, or Frrole, a big data enterprise that provides insights using Twitter threads, became active.

In the Gujarat Assembly elections of 2017, Inexture.com’s Mahipalsinh Rana created a platform for the state’s MLAs to add and manage party workers online, calling it Politician Digital Platform (PDP). “We are now taking it to the party level,” says the Ahmedabad-based Rana, adding that there is a crying need for last-mile data updates (read real-time) in India for AI to kick in. “When voter data gets updated, we can pinpoint in which area a candidate needs to work more,” he says.

Rana’s friend Saurabh Vyas, chief mentor of the Bengaluru- based Yobny Technologies, approached the Election Commission (EC) in February to present data from the top 15 Indian cities where only 5-18 per cent of the urban population voted—much less than their respective state averages. Simply put, the idea was to get urbanites to the polling booths. “During our field research, we learnt that long queues deterred people in urban areas to vote and that’s how Queue One came along,” he says. On the Queue One app, a voter provides his voter ID number and receives a position in the queue along with map navigation to reach the booth. As a test case, the EC allowed Vyas to roll out the technology in one of the booths of the South Bengaluru constituency when it went to polls in April.

In Bengaluru again, Amit Bansal runs Nextelection.com, which amplifies digital campaigns of candidates. Bansal has also released an ‘Election Promises Tracker’ for the electorate, wherein the 100- plus promises of the BJP’s 2014 manifesto were analysed threadbare, resulting in ‘46 promises broken, 37 inadequate progress, 15 adequate progress and six fulfilled’. He is now venturing into big data for sentiment analysis, demonstrating which policies or candidates have done better, or for that matter, how candidates are perceived by voters and on what parameters. He’s also looking at blockchain technology in which everything a candidate utters goes into a digital public ledger. “And since the ledger is auditable, we will have a blueprint of the reputation of candidates,” says Bansal.

Gurugram-based start-up Silver Push is also setting up a sentiment analysis barometer for the Congress party using keywords. It analyses the success of a campaign once it is rolled out and then shares feedback so that the party can tweak its campaign.

What Lal is to the AAP, Praveen Chakrabarty, who was earlier associated with the Aadhaar project, is to the Congress. As chairman of the party’s data analytics department, he recently claimed in The Times of India how the party has given data dockets to all candidates, which contain information on households, new voters, missing voters and local issues. In the report Chakrabarty claimed, ‘Instead of pushing down a single leader’s message using technology, we empower our party workers with data and technology. These workers pass on personalised or customised messages to voters.’

Party IT cell chiefs are also looking at ways to address basic issues, such as fluctuating networks. The AAP’s Ankit Lal has been scouting for a low-cost solution for dedicated network connections across regions. He could do with a 4-5 Mbps internet connection under Rs 1 lakh. “The limitation with hotspots today is that they are all mobile network-dependent. As soon as the number of people in an area rises above 2,000, the mobile network collapses. We want to use such networks to cover our rallies,” says Lal, claiming that national parties like the BJP and Congress rely on the more expensive satellite connections for their rallies and are billed per second.

He underlines the need for video compression saying if his team records a 30-minute video of a rally and an immediate internet connection is unavailable, sending a 4GB footage to him would be a huge effort. “Having a compression technique which gives the same quality while compressing the video well is vital… all this needs to be mobile-enabled using the Internet of Things,” he says.

He is optimistic about General Election 2024 when rural India will also come into the ambit of big data because of telecom advances like 5G. Also, with Tesla boss Elon Musk announcing the launch of more than 7,000 internet-beaming satellites and the US Federal Communications Commission approving it last year, Lal hopes costs would be low enough for his party to benefit. Similarly, Google’s parent Alphabet’s Loon project can also come into play soon. It uses high-altitude balloons at approximately 18 km to create an aerial wireless network ratcheting up 4G-LTE speeds.

In the Rajasthan elections last year, Karan Singh’s N-Sight Consulting worked with the Congress. In the Chomu Assembly seat, secondary data showed 100 per cent anti-incumbency. But when his team sifted through primary data, it showed that BJP’s Ram Lal Sharma, the incumbent, was both popular and had an active social media footprint compared to his Congress opponent Bhagwan Sahay Saini. With big data and social media analytics, Singh predicted Saini would lose by 1,000-2,000 votes. The result: Ram Lal Sharma won by a margin of 1,288 votes.

Again, take the case of Vivek Dhakar, the current Congress MLA from the Mandalgarh constituency in Bhilwara district of Rajasthan, who won in the 2018 bypoll. In many ways, Dhakar was machine-picked by Singh and his team who studied secondary data and then collected primary data across all the 144 villages in the region. Of the two-three candidates Singh’s client, the Congress, proposed, his team chose Dhakar because algorithms showed a strong base of committed voters. The team predicted a victory margin of 13,000 votes. Dhakar won by 12,976 votes.

As Singh puts it, “It’s all about precision. More than 2014, this election proves the prowess of technology to tilt the scales in favour of politicians who invest in it. Big data, AI, machine learning, virtual reality and more—you will see all of them come into play seamlessly in a connected world giving more bang for the buck in General Election 2024.”

Also Read
Other stories of General Election 2019

2019analysisGeneral Electioninformation technologypoliticstechnology