Open data cuts through Kerala election noise

Every election season in Kerala comes with its familiar rituals – predictions, projections, WhatsApp forwards flying in all directions. Even if you’re not someone who follows politics closely, there comes a moment when you try to remember: who won my constituency last time, what was the margin, and has anything really changed since then?

It sounds like something you should be able to check in a minute. The data exists. The Election Commission has it. Party websites publish their versions. But anyone who has tried to piece together voting patterns across Assembly, Lok Sabha and panchayat elections knows how quickly it overwhelms you. Browser tabs multiply, numbers don’t line up, and somewhere along the way, you give up.

This is the gap that OpenDataKerala’s Kerala Legislative Assembly Election 2026 Data Portal is trying to close.

It doesn’t generate new data. It pulls together what already exists – across Assembly, Parliament, and local body elections – and presents it in one place, organised by constituency, with charts and visualisations that actually tell you something. Each of Kerala’s 140 constituencies gets a structured snapshot, and then the real value – vote shares, margins, and how results have shifted across election cycles.

More usefully, the platform connects to an MLA track that logs what sitting legislators actually did during their term – questions raised in the Assembly, unstarred questions, participation records. So you’re not just looking at who won. You’re looking at what they did with it.

Useable data: “The primary objective is to ensure that reliable and structured information is available to a wide range of users, including the general public, political workers, researchers, and journalists,” the OpenDataKerala team notes.

For anyone comfortable with data, this goes a step further. Constituency-level information can be exported, shared, or even fed into analytical tools or AI platforms to surface patterns. The platform also integrates results from recent panchayat elections, which matter because local-level signals often show shifts before they become visible at the Assembly level.

The Open Data Kerala activists collected results from 941 grama panchayats, 86 municipalities, and 6 corporations in all fourteen districts into an interactive map. In an election season where opinions travel faster than facts, such efforts matter.

Global Initiative: The story of how this community came together goes back well before the current election cycle. Many volunteers had been involved in global forums like OpenStreetMap and similar efforts on platforms like GitHub. A small mapping problem in Koorachundu grama panchayat, Kozhikode, in 2014 gave the community its early momentum, says Jaisen Nedumpala.

He was working in the area when the Election Commission requested clear boundary data for a remote panchyath called Koorachundu in Kozhikode that straddled Koyilandi and Thamarasseri taluks. Official datasets were expensive and out of reach for a cash-strapped panchayat. Instead of waiting, he turned to open-source tools and communities like OpenStreetMap to build the required map.

“The task involved a lot of physical activity… it was a vast area with a thin population. Identifying and marking important buildings meant a lot of journeys. But the end result was very satisfying.”

The map was accepted. More importantly, it showed that a group of volunteers using publicly available tools could do what formal systems struggled to deliver.

Bigger Role: Over time, this network of volunteers began to make its presence felt. During the 2018 floods and later during the Covid-19 period, volunteers helped build dashboards that supported rescue and relief coordination. Naveen Francis, an IT professional associated with the community, recalls how existing official data often lacked critical details.

He cites the relief work during the Wayanad landslide in 2025. “By overlaying satellite information and other spatial data into the PDFs they had, we were able to quickly create a tool they could use to carry out the rescue and relief plans.”

Despite such examples, Kerala has had a chequered history when it comes to making data accessible. India has had an open data policy framework at the central level since 2012, but implementation at the state level has been uneven. In Kerala, discussions have moved slowly, often getting caught in process.

In the meantime, departments have continued to operate in silos. Data exists, but not always in compatible formats. In some cases, access is limited due to the absence of clear guidelines. For anyone trying to work with it – researchers, journalists, even government teams – the friction is real.

Keep Building: OpenDataKerala’s approach has been simple: don’t wait. Use what is already publicly available, aggregate it, clean it, and make it usable.

The 2026 election portal was built by pulling information from the Election Commission, verified party sources, and other authenticated datasets. The process itself became a learning exercise. Journalism students from Calicut University were brought in, giving them hands-on exposure to data-driven reporting.

“Some of them were unsure at first,” says Manoj Karingamadathil, who led the project. “But once they got into it, they started seeing how data can tell stories. By the end, they were proud to see their work being used as a public resource.”

The community now plans to take that further, offering structured training to help more students and professionals work with data. “Many shy away from it thinking you need coding skills,” Manoj says. “Collecting and studying data is something anyone can learn. The trick is presenting it in a way people can actually understand.”

Informed Voters: The 2026 assembly election platform is careful about what it claims. It doesn’t try to predict outcomes. Elections are shaped by too many variables – campaign narratives, alliances, candidate profiles, turnout, even the weather on polling day. Better-resourced forecasters have got it wrong often enough.

But the data does something more durable than prediction. It makes the past legible.

When you can see what actually happened in your constituency across multiple election cycles, you stop relying on loud voices and forwarded messages to form your picture. You start asking different questions – better ones.

That, more than any forecast, is what good election data infrastructure produces: voters who are no longer relying on noise, but on a record they can actually read.

Startups face a cash crunch

New rules set down by SEBI (Securities and Exchange Board of India) are choking angel investment platforms, as the norms have raised the bar for investors. Angel investments in Indian startups declined sharply in 2025, falling 44 percent to 834 deals from 1,495 deals in 2024, reports Moneycontrol. To qualify as an accredited investor, individuals must have an annual income of 2 crore rupees or more, or a net worth of 7.5 crore rupees, with at least 3.75 crore rupees in financial assets. Many existing angel networks may have to shut shop, dealing a huge blow to the early-stage funding ecosystem.

Funding by venture capital firms is also drying up, and the ongoing war in the Middle East is adding another layer of vulnerability. Indian startup funding showed a 17 percent decline, raising only 10.5 billion US dollars in 2025, down from 12.7 billion US dollars in 2024, according to the India Tech Annual Funding Report 2025 by Tracxn. It looks like conditions are tightening further. According to YourStory, total funding for the first week of April came in at 117 million US dollars from 20 transactions, compared to 328 million US dollars from 22 deals in the previous week.This is the third lowest amount raised by Indian startups on a weekly basis.

NeuroPixel.AI calls it a day

Startups offering AI-led services are finding it hard to sustain, and the latest to join the list is NeuroPixel.AI. The Bengaluru-based startup, which had raised around 1.2 million US dollars from investors, is shutting down its operations after about six years. Founded in 2020 by Aravind Venugopal Nair and Amritendu Mukherjee, NeuroPixel.AI developed AI-led solutions for the fashion ecommerce sector, including virtual try-ons, synthetic model generation and cataloguing tools. According to Nair, the company’s technology delivered comparable output quality at lower costs, but struggled to scale due to distribution challenges. It also came under financial pressure after losing a major client, with dues remaining unpaid for over six months.

Monarch Tractor runs out of road

Autonomous tractors were once a much-hyped segment, but US-based Monarch Tractor appears to have reached the end of the line. Once valued at over half a billion dollars and buoyed by more than 240 million US dollars in initial funding, the company drew major attention – in 2023, Time called the vehicle one of the year’s greatest inventions, and Forbes predicted it would become the world’s next billion-dollar startup. But the company is now preparing to shut down. There are several reasons at play, but chief among them is that the AI-driven tractors seemingly did not deliver on their promise.

Lost in translation

More bitching about AI, this time from South Korea. The KBS TV YouTube channel was live streaming the launch of Nasa’s Artemis II mission and had used AI for translation. When mission control said “Roger, roll, pitch,” the AI-generated subtitles completely missed the aerospace context, reading “pitch” as an English expletive beginning with the letter B – and helpfully translating it into Korean, reports Korea Herald. KBS apologised after viewers called out the mistranslation and quickly fixed the subtitles. A reminder that AI can still go off-script spectacularly – even when it’s just supposed to just translate.