The Indian think-tank National Institution for Transforming India, (NITI Aayog), an initiative by the Indian government has played a phenomenal role in advancing emerging technologies in India and putting in a roadmap for AI. Analytics India Magazine got in touch with the head of data analytics cell at NITI Aayog, Avik Sarkar, who uses smart data and advanced analytics technologies to help government, citizens and clients improve outcomes and optimise processes. With core strengths in analytics, statistics, machine learning & data mining, he is working across various industries to bring the best out of them. He shares his thoughts on big data agenda in policy making, enabling centralised data storage, and more.
Analytics India Magazine: What does your role at NITI Aayog involve?
Avik Sarkar: While government consists of several ministries and each ministry have their own divisions and pockets on analysing data, I head the analytics cell of NITI Aayog. We take a lot of data from outside, analyse it and try to make sense of it for economic good. That is our primary goal. There are different verticals such as healthcare, education, energy, science and technology, on which we are working based on the data requirements. We see what kind of visualisation can be done on this data that adds value. It is more of an internal role trying to fill out the data analytics gap in the organisation.
AIM: What are the different data sources?
AS: It can be raw data coming from states or districts, survey responses, survey estimates, and more. If you are looking at the status of healthcare of the whole country, you can look at hospital records which gives you a snapshot of all the patient-related data. If you have to look at the health of children of certain age group, we go into surveys by going to houses and take samples from there. We take a measurement of people and report back while conducting root cause analysis.
AIM: How is NITI Aayog pushing the big data agenda in policy making?
AS: Big data can mean different things to different people. When we talk about operational data, it has been captured over years and is being used for policy making. We try to look at big data to see how the formal and informal employment sector works, such as how many chartered accountants are coming out each year and how many jobs they are generating. On similar lines, we try to look at retail data and try to understand how the economy is moving on several fronts. We also have an aspirational district program where we have a dashboard for monitoring the data on the developing indicators in these districts in real time. The district collectors are filling up this data which is uploaded in real time in champions.change.org and then we look into that data to check performance. Some can be sectorial viewing of districts; some can be about jobs in a sector for the economy. It depends on the type of activity and we employ different methodologies for that.
AIM: Would you talk about areas where NITI Aayog is shaping policymaking in digital payments, healthcare or agriculture?
AS: Particularly nutrition is a big area here. National Nutrition Mission which is done in collaboration with the Ministry of Health, Ministry of Child and Women Healthcare Development. And, NITI Aayog also has a technical evaluation for the National Nutrition Mission. We are monitoring how the program is running by evaluating the output. Our other key focus areas are agriculture, healthcare and education. These are the strategically important areas from a social perspective. We look at a lot of satellite imagery and AI-based technologies to improve the agricultural yield.
AIM: Would like to talk about your work in energy vertical. Energy sector modelling, energy data management?
AS: This started in 2016. Healthcare and education data analysis are short-term in nature, which might take 1-5 years. But in case of energy, we are looking at the future horizon, say 20-30 years. New energy generation plants are coming up and we do energy forecasting for the future of supply and demand. We have to be very strategic in nature. For instance, when we import oil, whether the oil price will increase in the future or not, needs energy forecasting. This decision can impact whether the government recommends the move towards electric vehicles. While we have a surplus of electricity from wind and solar energy, we say there will be electric vehicles 10 years from now—this is a strategic move. It will take some time for these things to percolate in an economy. The industry will take some time to grow. The industry will take some time to grow. The competition will evolve and more cars will start to appear in the electric vehicle domain. The policy needs the industry to grow. Infrastructural changes need private players to participate and things like this take time.
AIM: How are you strengthening centralised data storage?
AS: We have a project for that, which is the National Data Analytics portal. There is a lot of data being put by ministries on their own websites. We do not have a central repository for all this data. So, with this portal, we can have all this data accumulated on a central portal which will be accessible to all the people across different sectors. This can be used for policymaking.
AIM: There was news that NITI Aayog is looking to hire more people. Are these positions related to AI, Data Science?
AS: It is divided into verticals such as subject matter experts on healthcare, education, agriculture etc., where most of the hiring happens. Data analytics and IT are more horizontal roles.
AIM: Would like to talk about technology stack at NITI Aayog? What are the tools that you use in Analytics?
AS: We internally use tools like Excel and other proprietary tools for charting and mathematical analysis. We have large computer infrastructure and tie-ups with institutions such as IIT Delhi for its supercomputing facility.
AIM: What are the challenges that you see in the space?
AS: We aim to use data for economic growth and innovating various activities in the domain. One of the challenges for the government is how to get access to this private data. There are retail chains which are using it for customer analytics. If we can combine this data and form a consortium, then it can be used to analyse retail trends across the whole country. This will not hamper any competition but will help policymakers in understanding how the consumption in certain areas. Also, we can look at telecom data from telecom players and other data which is lying in the domain to be analysed. We need more real-time data either from the government or corporate and make sense of the data for social good.
Another challenge is about educating people about what data can do for you. Once they are educated, they will start asking the right questions and go beyond counting and reporting aggregate numbers.