A prospective jewel market investor wants to know if the limit has already been reached for opening new stores at a particular location in a city. The government of that same city also needs to come up with a plan to install hospital equipment and needs to know which regions are most in need. These and countless other decisions, whether involving investments or public services, can no longer be obtained with accuracy without the combination of a series of elements processed by highly precise instruments. This is where Data Science comes in, which is already a reality in the most innovative educational institutions.

FGV’s Sao Paulo School of Economics (EESP) brought together leading experts on the subject in the 1^st Brazilian Data Science Meeting, held at FGV’s 9 de Julho Auditorium on December 10-11. Eduardo Francisco, a professor from the Department of Information Technology and Quantitative Methods of FGV’s Sao Paulo School of Business Administration (EAESP), opened the meeting by analyzing the state of the art of the use of Data Science in the geospatial industry.

For the specialist, the use of other non-income variables can be valuable to establish a set of more precise information, determining factors for decisions in both the public and private sectors. One of the most recurring is electric energy consumption, which has a high correlation to income.

“The great breakthrough of recent studies is the replacement of income analysis with energy consumption without any prejudice to the accuracy of the information, with the advantage that asking about a citizen’s energy consumption is less privacy-invading than asking about their income”, said the professor.

Data Science does not only apply to investments. Emerson Marçal, a professor from EESP, talked a bit about inflation forecasts using Data Science instruments. For Marçal, it is possible to use existing techniques to improve those forecasts.

“We were invited by Central Bank technical experts to introduce new techniques. The processes generated are based on unstable data with possibilities of breakage, and when we began to discover the peculiarities of the scenario, everything changed. In addition, a model based on data science can avoid some imprecision in the results. An analysis based on the historical cycle, for example, can point to a recovery when it actually does not exist”, he said.

Other examples of uses for Data Science were mentioned, such as humanitarian actions, logistics planning for the highways of the State of Sao Paulo, retail planning, news classification and press reliability, the healthcare industry and other topics.