EDITIONS
Stories

How big data and predictive analytics could make Kumbh Mela and other large scale events safer

Harshith Mallya
1st Aug 2016
Add to
Shares
230
Comments
Share This
Add to
Shares
230
Comments
Share

Modern technology has improved our lives for the better. Recent advances in technology have shown that some of our science fiction fantasies could indeed soon be realities. The popular 2002 science fiction flick Minority Report explored the possibilities of technologies that could help law enforcement capture criminals even before they committed crimes.

We haven’t reached that ‘utopian stage’ yet. However, advances in big data and machine learning have shown that, with a big enough data set and predictive analytics, it is possible to predict human behaviour and take corrective action in advance for some use cases.

YourStory-kumbh-Mela-3

Image credit- Shutterstock

A recent collaboration by an Indo-Dutch consortium that included the Government of India, the Netherlands organization for Scientific Research (NWO) - the Dutch R&D organization participants from organisations including IISC, Bangalore, IIT Kanpur, and two startups - Tarah Technologies and Aarav Unmanned Systems looked to explore the possibilities of this by collecting data from the Kumbh Mela that was held recently. Here is a story of the whole initiative and how a bootstrapped Tarah Technologies became a part of it.

Story so far - From CRM to big data

Neelim-Tarah-tech
Neelima Vobugari

Started in late 2010, Tarah Technologies is a big data consultancy firm focussed on competency development and consulting. The startup originally started out as a customer relationship management (CRM) consultancy firm and in 2014 pivoted to operating in the big data domain prominently focussing on machine learning.

Neelima Vobugari, Founder of Tarah Technologies, is a Big Data Consultant and has nearly 16 years of experience working in different IT domains. Before starting up, she worked at multiple startups and companies including IBM for almost a decade. She then felt a strong pull towards entrepreneurship and starting up on her own. So she equipped herself with the required skills through a course on entrepreneurship for women entrepreneurs at IIM Bangalore in 2009.

Srinivas
Dr Srinivas Padmanabhuni

At present, the startup has around 10 employees with varying skills, with primary among them being in big data and machine learning. Dr Srinivas Padmanabhuni functions as the Chief Mentor and brings to the table his experience of about 15 years in the IT sector. He has a Ph.D. in Artificial Intelligence from University of Alberta, Edmonton, Canada, and also has B.Tech and M.Tech degrees in Computer Science from IIT Kanpur and IIT Bombay, respectively.

Prior to being part of Tarah Technologies, Srinivas was Associate VP heading research at Infosys till October 2015 and the President of ACM India till June 2016. He also has seven granted patents, 15 filed patents, one published book by Wiley, and over 70 refereed journal and conference papers to his credit, in addition to invited talks and editorial positions.

Having bootstrapped so far, Tarah’s two main offerings to generate revenue include consulting and training related to big data for different clients. Tarah declined to name their clients for confidentiality reasons, but added that they work with large educational institutions which contribute to the skill development of individuals across different domains like Business and IoT analytics. Neelima said,

In our consultancy, we use machine learning software to solve real world problems from the data collected.

The platform had provided training services to six clients for around six months. Despite the demand for training programmes, they decided to put it on hold to focus more on their consulting business. Neelima noted that they have consulted with five clients over the past year and sees this sector growing though it is still in its infancy in India.

Tarah works on a flexible project or monthly billing model based on the needs of the clients. Based on their expertise and having worked with different clients in the past and also because of Srinivas’ extended network, Tarah Technologies was recently approached and chosen in an Indo-Dutch consortium of a team of companies, academicians, and researchers from different domains in an effort to automate crowd management at the Kumbh Mela.

The long-term aim of the ‘Kumbh Mela experiment' is to develop advanced methods and algorithms to support events planners and managers manage extremely large groups. This project is being funded by NWO and Government of India’s Department of Electronics and Information Technology (DIETY). Neelima noted that considering the initiative was a noble cause with far reaching consequences, Tarah had offered their expertise and platform (servers, etc.) pro-bono for the cause. DIETY and NWO sponsored travel and other related expenses.


Related read from May 2016: Startups and govt make Ujjain Simhasth the most high-tech Kumbh ever


Big data and Kumbh Mela

It is a well-known fact that during large scale events like Kumbh Mela, Hajj Yatra and other similar events, the footfall of crowds is in tens of millions. So managing such unexpected crowds is difficult in normal scenarios, along with the enhanced risk of disasters such as stampedes occurring due to confusion.

So, to better understand the underlying pain point and solve this issue, different organisations part of this multi-year multidisciplinary research effort are working on three key research areas related to the understanding and management of massive human crowds - data collection, data analysis, and modeling/prediction.

Kumbh-Mela-experiment2
Srinivas Padmanabhuni and Neelima Vobugari

The data collection phase happened during the recent Kumbh Mela that occurred from April 22 to May 21 at Ujjain. As Kumbh Mela is considered among the largest such gathering in human endeavor, the team of researchers and companies felt that it was the ideal project to gather a large volume of data for diverse model building. The data were collected through a variety of means via sensors, go-pro cameras, drones, apps, physical questionnaires, videos, call data records, etc., all of which were fed to big data analytics to enable derivation of models. Neelima noted,

Patterns of movement of the crowd is for the first time being studied anywhere in such a comprehensive manner with real time diagnostics. When a situation like a stampede can be predicted well in advance, precautionary measures are taken to avoid life loss.

Apart from Tarah Technologies, Aarav Unmanned Systems (AUS) was the other Indian startup that was part of this initiative. With the help of AUS’s drones, the Kumbh Mela team was able to effectively model the entire layout of the Kumbh Mela venue at Ujjain. Yeshwanth Reddy, Co-founder of AUS, noted they had also planned to deploy drones during the event to collect more data. But, because of some regulations that cropped up at the last moment, they were not allowed to. But ground-level data were collected through various other means.


Related read: Aarav Unmanned Systems is mapping the future with its UAV technology


Post the data analyses and modelling stage, which will take some time because of the large volume of data, the aim is to create a decision support software that predicts the crowd well in advance and alerts the organisers in form of an early warning system.

Interesting implications

While the current use case is for Kumbh Mela, Neelima believes that this ‘crowd management system’, which relies on research from machine learning, complex systems, and organisational behaviour, could be tweaked to other use cases such as managing traffic on roads, pedestrians on busy streets, and tourist areas with large footfalls.

As we move into a more connected world where buzzwords like internet of things (IoT) and connected devices are slowly becoming mainstream, it will be interesting to see what more can be done with these devices and sensors that track a lot of useful ground level data.

Website- Tarah Technologies

Report an issue
Add to
Shares
230
Comments
Share This
Add to
Shares
230
Comments
Share
Authors

Related Tags