“In terms of congestion, Bangalore is one of India’s worst cities,” says Dr. Raghuram Krishnapuram, a senior scientist at the Centre of Data for Public Good (CDPG), Foundation for Science Innovation and Development (FSID), in Indian Institute of Science (IISc) Bengaluru.
The traffic jams are not new to the city and are as common here as its climate and start-up environment with its green parks. In 2022, Bengaluru was rated as the second most congested city globally by TomTom Traffic Index, only behind London for that infamous position. And while this has improved – Bengaluru dropped to 6th on this list in 2024 – it continues to be one of the country’s most congested cities, something any Bengalurean will attest to.
CDPG and the Bengaluru Traffic Police (BTP) have partnered to organize The Bengaluru Mobility Challenge hackathon aimed at coming up with AI/Geo-spatial solutions to address traffic menace in the city. “There is a huge potential from my perspective to make a difference in Bangalore”, Krishnapuram claimed.
Data-driven solutions
The incubation centre called FSID that used to be called SID (Society for Innovation and Development) exists at IISc so that IISc scientific knowledge can be harnessed for societal benefits and industry needs. One such initiative was India Urban Data Exchange (IUDX), which started here. To enable data sharing between municipalities’ agencies, private sector businesses, citizens and other government enterprises; it seeks to eventually help cities use this data toward resolving complex urban issues.
On November 24th, 2023 when CDPG formed under FSID, it absorbed IUDX hence becoming part of it. According to Krishnapuram: “The idea is that cities in India can leverage the data that they have to do impactful things.” He goes on to say, “You can derive a lot of insights from the data and enable an entire ecosystem of start-ups that can use that data to build useful applications that will benefit society.”
Krishnapuram explained how Tripura’s capital, Agartala has partnered with IUDX to demonstrate how data can be applied for public good. Video data from several different intersections in the city was collected in late 2022 for traffic monitoring and control purposes. “For example, if there are a lot of vehicles accumulating in one direction, you can increase the duration of the green light, etc.,” he says. Because all these roads are interconnected, it is necessary to coordinate a lot so as not cause any sort of traffic disruption.
While this pilot project did not go beyond the stage described above, many more such projects are ongoing. It is already running in almost fifty Indian cities and being utilized for various other reasons too. For example, in Surat IUDX is used for monitoring the bus system while Chennai utilizes it to assess water levels in reservoirs and predict flooding. “Also being used to setup green corridors and solid waste disposal schedules… those kind of things”, he adds.
The Origin of the Hackathon
One speaker at a symposium during the launch of CDPG in November 2023, M.N. Anucheth, Joint Commissioner of Police Traffic Bengaluru City spoke about the traffic situation in the city and requested that entrepreneurs assist in solving it. “So, we thought that maybe this would be a good problem to work on,” recalls Krishnapuram.
That led to The Bengaluru Mobility Challenge that was announced earlier this month. The participants will rely on camera data provided by BTP. “Instead of just using the BTP cameras which are limited in number, we are leveraging the Safe City cameras to look at how many vehicles are passing in large parts of the city,” he says. If one can do AI-based analytics, this data can be used to model the city traffic, he adds. “Bangalore is a big city but it doesn’t have any traffic model; nobody really knows where people travel now.”
Krishnapuram believes that insights from data could assist with city planning and decision making such as expanding public transport system, designing alternative routes, deciding which roads should be one-ways or putting a metro line.” If you have this information you can do lots of things that make it easier for people to commute.”
Two Phases
The hackathon will consist of two phases open for all residents in India focusing on different aspects of urban mobility management. The first phase will involve using Safe City installed around IISc cameras’ data so as to predict vehicle counts and generate traffic patterns. Understanding these patterns can help predict traffic, which in turn can help police take action,” he says.“This could help them regulate the traffic differently.For instance close certain roads to divert the traffic.”
Whereas Phase Two will see participants re-identify vehicles across different network locations and estimate origin-destination flows. This will help understand motorists’ behavior within the city. Krishnapuram recognizes that this is not a simple task given the large number of vehicles that look alike. However, he hopes that data from this phase would help them to test their hypotheses about driving behaviors.
Hackathon teams can have up to a maximum of five members and decide to participate in one or both phases. He says, “people don’t need to attend phase one in order to be able to attend the second phase.” Conversely, if a team wishes to stop after the first part they may do so but they are required to make their code public by putting it up as open source. If these results are followed through on by those who participate in Phase Two that is okay. On the other hand though, anyone else who proceeds into Phase Two does not need this.
Feasibility and beyond
‘If things go well with the challenge,’ he hopes ‘then we hope to scale it up to the whole city.’ But currently, we are focusing on feasibility aspect, he says. “We want to see if this data is useful…if you can do something with it using AI and machine learning.” For instance; Indian traffic is just described by him as “a different beast that is completely chaotic” than existing AI algorithms for traffic developed using data from other countries that never really worked here. Take for example vehicle count algorithms fail in India owing to our roads being happily trundling along buses, cars and motorbikes with rickshaws and bullock carts too. “They don’t get noted down at all,” he notes further.“That is a challenge.”
Krishnapuram sees a possibility of Indian specific traffic solutions coming out of this hackathon. He stated that ‘when you throw open such problems unto many people then everyone comes up with different ideas’. Additionally since these solutions are open-source he is expecting that some of the solutions may come out as winners during hackathons. Which could also be utilized for entire cities or even countries like India.“Traffic is a huge problem in India, and we have to find sustainable solutions,” he says.
Beyond traffic issues
But no! The use of data to address civic issues meaningfully is not just limited to traffic matters alone. “Traffic and transportation are one point, pollution and flooding are other issues. We hope that through hackathons awareness will be created and good solutions.”
