The start-up of the day: Datascience, data scientist challenges at the service of companies
Axa and SNCF have already used his services. But the start-up has to face the reluctance of some companies to make their data available to a large public
Frenchweb invites you today to discover Datascience, a Parisian start-up founded in 2013 whose objective is to bring together a community of data scientists to participate in challenges serving companies. More details with Arnaud Laroche, the co-founder.
Frenchweb: How did you get the idea for your company?
Arnaud Laroche: Datascience.net was born in November 2013, from the common will of Bluestone, our consulting company specializing in the treatment of Big Data since 1996, and GENES (Group of National Schools of Economics and Statistics, composed in particular of Ensae-Paristech, Ensai and the Secure Data Access Center), to create and lead the community of Data Scientists, in order to create a bridge with the business world around issues oriented Data Science.
The Data Science, which involves creating knowledge from the study of volumetric data, is a relatively new discipline. Businesses have growing needs in this area. However, France benefits from a real pool of experts, thanks in particular to the quality of its training in mathematics and its strong scientific culture. But all these talents and all these needs are still struggling to meet.
We wanted to remedy this because we were aware of the solid lever for development that the community can represent and its desire to put its expertise into practice. It is therefore quite naturally that Bluestone, approached GENES, to jointly develop the platform. datascience.net, and involve all stakeholders in the field.
What need are you responding to?
Datascience.net meets the needs for innovation, improvement and adaptation around data in companies. However, organizations often suffer from a lack of resources. Our platform allows them to meet their objectives by drastically limiting the consumption of their resources: time, people, expertise and financial.
Datascience.net allows companies to submit to the community Data Scientists, scientific problems of exploiting big data, and thus bring about completely new solutions.
For example, how can services be improved at SNCF stations in Ile-de-France, by making precise estimates and forecasts of attendance at each of the stations from multiple available data sources?
For companies, it is an opportunity to test new models or new approaches, complementary to the work sometimes already carried out in-house, and to find new ideas, new technologies or new algorithms, ultimately to find new, more effective solutions brought to you by the power of an entire community.
For the participants in the challenges, beyond the obvious interest of the bonuses offered to the winners, it is also an opportunity to confront their peers, to measure the effectiveness of their approaches and to discover new methods or new tools for other participants. It is also obviously a framework which offers a certain visibility to their professional profile, for the best of them. Today, datascience.net brings together the best Data Scientists French-speaking people, and allows you to participate in exciting challenges, for large accounts such as AXA or SNCF.
Very simply, how do you make money?
The economic model is based on fees invoiced to companies offering challenges (to cover the work of preparing and implementing the challenges), and on a commission deducted from the amount of bonuses involved.
Ultimately, our database of Data Scientists will constitute a complementary source of income, by offering recruitment assistance services for companies looking for highly targeted talent.
Who are your direct or indirect competitors?
One of the main platforms that offers the same kind of scientific challenges is Kaggle, an Anglo-Saxon start-up, which addresses the international community.
The approach of datascience.net is however different from that of Kaggle, in the sense that we have the firm conviction that companies, at least in France, need support, and people who are close to them.
The specificity of our platform: offering companies support in their procedures, and guaranteeing a barrier-free dialogue with Data Scientists who submit the best solutions, the one they will implement … The operational, physical relationship therefore seems fundamental to us. This does not ultimately prevent a more international development of the platform, but with a concern to be a local player each time.
What or which companies are you being compared to in error?
The concept of datascience.net is quite innovative and quite original in the field of data science, while remaining very simple and easy to understand. We are therefore rarely wrongly compared with other services.
What was one of the first issues in your development, and how did you deal with it?
One of the first difficulties we encountered was the reluctance of some companies to make their data available to a large public. This obstacle, quite natural, is however quickly overcome by exchanges and by pedagogy: the openness of data and the value created by knowledge is often greater than the supposed risk of losing a competitive advantage by creating a blockade around its data. . This has long been demonstrated by the open-source movement and the more recent open-source movement.data.
This obstacle can be overcome all the more easily as many solutions can be found, such as the partial masking of certain sensitive data or their anonymization. Datascience.net has also set up private challenges, a service that allows companies to select the Data Scientists who can access their data.
Finally, for challenges that rely on really very sensitive data, for which all the solutions mentioned would not be enough, our partnership with the Secure Data Access Center (CASD) is a strategic advantage. The CASD offers a technology, hitherto reserved for research, making it possible to guarantee absolute data confidentiality. This technology has been approved by INSEE to provide research centers with access to personal data from its surveys.
What is your main asset in this market?
Datascience.net is an easy-to-access, simple and fun platform that offers a unique playground for Data Scientists French, and which offers companies access to this community and to the power of collaborative intelligence… So the provision of a lever for technological development. We establish a win / win relationship for the different players.
What is the best advice you have been given and by whom?
“Only do what you really believe in.” Believe in what you do. And do it in your image ”- Family Council.
Who is the personality you admire the most?
Elon Musk is a fascinating character. After reselling Paypal which made him a multimillionaire, he created SpaceX to send tourists into space, then Tesla a new brand of revolutionary cars, Hyperloop magnetic capsules on air cousin to transport people faster than by plane… His ideas seem crazy every time, he sets no limits, and it works. His first rockets supply the ISS station on behalf of NASA, Tesla is a huge success. A brilliant visionary entrepreneur repeat offender.
Founder: Datascience was created by Bluestone, a consulting company specializing in the treatment of Big Data, and the GENES (Group of National Schools of Economics and Statistics) made up in particular of Ensae-Paristech, Ensai and the Secure Data Access Center (CASD).
Community: 470 data scientists one month after the official launch of the platform, pseveral hundred contributions per challenge
Investors: tAll investments are made on their own by the founders
Creation date : 2013
Company based in: Paris