DATA SCIENCE FOR EDUCATION IN INDIA
Clémentine Vignault Rao
A former masters student in Cambridge, Clémentine Vignault Rao is now the co-founder and CEO of a Cambridge-based start-up, Slate2Learn. Incubated at the Cambridge Social Ventures at the Judge Business School, Slate2Learn is a micro-franchise of tablet-based tutoring centers. They opened their first three centres in January and February this year in Delhi.
Building on her earlier business experience in India and France, Slate2learn data analytics capabilities are largely a result of Clémentine’s self-taught coding skills. We asked Clémentine about what led to her starting her own company, how she learnt the necessary technical skills and what lessons and advice she would pass onto other young entrepreneurs.
Clémentine, please explain the idea behind Slate2Learn.
Slate2Learn tackles the problem of low-quality education in India, through technology.
There are a lot of great learning technologies around the world, but for various reasons, they end up not being usable by people at the bottom of the pyramid in India. The challenge Slate2Learn addresses is: ‘How can we deploy learning technologies effectively at the bottom of the pyramid in India?’
To solve this problem we have created a data-driven, affordable and turnkey learning environment that we scale through a tuition centre micro-franchise model. Plugging ourselves in the massive after-school tuition market in India, we offer existing or aspiring tutors the opportunity to start a Slate2Learn tuition centre, and provide skills to up to 100 children in their community.
These entrepreneurs receive a ‘tuition centre in a box’ that consists in 10 low-cost android tablets, and a raspberry pi server, (equipped with Slate2Learn adaptive learning system and learning management system), and can run the centre in their home. Slate2Learn takes a percentage of each centre’s earnings.
This sounds like an ambitious project, what was your motivation behind starting Slate2Learn?
While living and working in India, I’ve observed first hand the problems that cripple India’s education system. I was very impressed by what some ventures such as Bridge International Academy had been able to achieve in Africa to tackle these issues. At a time when hardware was becoming more and more affordable in India, I thought there was a unique opportunity to use technology to bring new solutions to these old problems. I turned the problem in my head for almost a year and I finally decided to jump off the cliff, and I started Slate2Learn in 2015.
Despite being so young, Slate2learn has already come to the attention of Cambridge's Judge Business School. What is it about Slate2learn that is new?
The first difference between Slate2Learn and most other tech-based educational projects in developing countries is our business model. Even though our price point is extremely low (~£3 a month), Slate2Learn micro-businesses (tech-enabled tuition centres) are financially self-sufficient, and do not reply on donations to keep running.
We launched our first centre in January, in Sangam Vihar, a suburb of Delhi, and it broke even in the first week! We are clearly fulfilling a demand in this market. The second difference is how we use data at every level of our model, to reduce operational costs and maintain high learning outcomes at scale.
Can you explain in more detail what type of data analysis methods you implement and how this is used in the Slate2Learn model?
We use data at many levels of our model:
1) To tailor each child’s learning experience, the tutoring system uses data analytics to determine parameters used in the learning algorithms. Once we collect enough data, we will use machine learning to elucidate the different parameters.
2) In the classroom, data is used in a more ‘classic way’ to provide tutors and parents insight about what children are doing. We have a monitoring screen for tutors to visualise the classroom and alert systems for the tutor to intervene when required. There is also a reporting system for parents to keep track with their child’s learning progression.
3) We have a cloud-based dashboard on which we can visualise how each of the centres perform on a series of metrics. Here we use more business intelligence technologies. We use this information to maintain quality of learning across centres and to streamline various business processes.
On this topic, we notice you previously worked in geospatial optimisation in India. This seems to be a popular and fast growing area. What are the future challenges in ‘geo-data’, in your opinion?
A lot of decisions could be improved by integrating geo-data into the problem. There is a lot of scope for improvements in geographical datasets for decision making, but also in geographical data processing (like geocoding) although these areas have experienced fantastic improvements in a past 5 years.
Many of our readers will be scientists who are interested in moving into data science but perhaps have little experience, how much data science or data analysis had you done prior to embarking on Slate2Learn?
Prior to Slate2Learn, I had very little experience with data science. I have an engineering background, and thanks to this I have been able to pick up various technologies quite quickly. I think the most important thing in data science projects is to think through the problem you are trying to solve and really understand what you want to achieve with data. Mastering the different data-analytics techniques can follow naturally with experience thanks to the wealth of documentation and help available online.
It sounds like you have learnt quickly. How have you enjoyed these early stages of starting a company?
I have really enjoyed the freedom of building a project straight from my imagination. But the truth is, it is also very lonely at times, and as an entrepreneur you often have to ‘cross the desert’ for a few years before getting any recognition for your achievements, and there is plenty of time for doubt.
There could be some parallel with challenges young scientists face here. How do you deal with self-doubt? What advice would you give to budding young entrepreneurs?
I think doubt creeps in, and I would say this is especially true for women, when you look at what people do around you, and you tend to overestimate their achievements, compared to your own. It’s a bit the same psychological effect as when you look at people’s vacation photos on facebook. People will post a picture of a stunning landscape, with no one else around, when in fact there were 1000s of other tourists around but you just can’t see them on the picture. You have to be aware that most people will paint the shiniest possible picture of what they are doing. So I think one of the best way of combatting self-doubt is to build awareness of this bias. You can also build this bright sunny picture for what you are doing, and always keep it in a corner of your mind.
Question what you are doing, and what other people are doing, but most importantly don’t give up!
In my opinion, perseverance and focus are the most important skills for an entrepreneur. Perseverance because the road is long for a first-time entrepreneur, and focus because there are a thousand ways of getting distracted along the road.
There is a strong theme of environmental studies in your past experience (Clémentine has completed a masters in environmental policy at The University of Cambridge and worked at CIRAD, a public research institution specialising in agricultural research), why have you now decided to move into a social venture?
I spent several years exploring the interplay between environmental and development issues, in particular in the context of agriculture. Little by little I came to realise that the social angle to this problem was the most motivating for me, so I naturally shifted toward projects that generated more social impact.
We understand your lessons are currently tailored for primary school age. What subjects are most important to master for children at this age?
Numeracy, and in particular mental math, is very important as it is a prerequisite to a lot of other learning. In India, English language is very important, because it opens the door to the formal economy, and high paying jobs. Inequality of access to English language education leads to huge inequality of opportunity in India.
In addition to these subjects, considering the rising importance of being ‘data-literature’, do you think schools should include data science as a subject?
Yes, absolutely. Being capable of accurately thinking through a problem using data is a very important skill at a time when we see more and more use of data to prove false claims via ‘incorrect’ data reasonings.
Finally, what is latest success for Slate2Learn and where do you hope it will be in 10 years time?
As I was mentioning earlier, we opened 3 centres in Delhi in the past 2 months, and we got a very good response from parents and tutors. This is a very encouraging result given that what we do (tablet-based learning) is very new to our socio-economic target group. In 10 years, we hope to have opened 100,000 centres in India, catering to 10 million children, and that by then our technology will also be used in government schools as well.