Making the data talk, making sense of it and transforming it into business solutions – here are some of the miracles achieved every day by data scientists. An expert paints a portrait for us of this highly sought-after expertise.
Gaëlle Ramboanasolo is a data analyst for the cultural sector. No one knows better than she does how to analyze theatre audiences, from data collected by the ticket offices, for example. Spurred on by what she sees as a “real new paradigm”, she is currently setting up her own company in artificial intelligence and advanced analytics serving creative industries. “It’s not just a matter of analyzing the data,” she says, “but also developing algorithms and modelling and visualization tools that can be used independently by our customers.”
According to a report published by IBM in 2017, job offers for data scientists increased by 5% during 2016 alone in the United States, and competitive salaries range from $105,000 to $117,000 per year.
“Data is the new black gold!”, exclaims Gaëlle Ramboanasolo. “It can be exploited to solve all kinds of problems. In culture or in business, it brings a new philosophy in customer relationships. In other contexts, particularly in the environment, visual data analysis (such as satellite images of the territory) will one day help secure populations at risk.
For the data scientist the days follow each other and are not the same. But the work nevertheless follows a marked out path, a route to follow to get the most out of a data set. The scientist has an interest in being a good communicator. The first step is always to explain the potential of the data to the customer, which he often is unaware of, then to identify his needs, in a kind of “strategy discussion”.
“Then we proceed with what is called the data diagnosis, a careful identification of the database content, and it is cleaned somewhat to clarify it more,” says Gaëlle Ramboanasolo.
The scientist will then “prepare a model of analysis and calculation from a specific question” to query the data.
A high-level computer scientist, the data scientist is often self-taught, the entrepreneur says. The profession attracts as many computer science PhD students with good analytical abilities as inventive enthusiasts who have learned on the job. “University programs obviously provide a good structure, but they are always a bit behind the times compared to the market. I nonetheless recommend a Master’s degree in business intelligence or data science,” concludes Gaëlle Ramboanasolo.