Work experiences

Data Scientist I in Travel Partner Group division

Research fellow under CIFRE PhD program

▶︎ Public transportation field generates a large volume of data, that are still insufficiently exploited. For example, ticketing systems collect a lot of information each time a smart card is validated. Vending machines although record every transaction made by a customer.

▶︎ Projects:

• Clustering passenger temporal profile

Transport authorities often use their sales data to segment their customers. The aim of this project was to find a new way to cluster a network passengers by answering this simple question: "Who travels when?".

▻ Creation of a new clustering algorithm: NMF-EM

▻ Use of open-data (demographic and socio-economic) to better understand the passengers behavior

▻ One proceeding admitted in an international conference and one preprint submitted to an international journal

▻ Development of a R package: nmfem

• Forecasting the public transportation attendance on a network

In order to improve the quality of an urban transportation network, it is very important to be able to anticipate the users' demand. As population travels are complex phenomenons dependent on a large number of variables, it is simpler to have a range of possibles check-ins than a punctual forecast.

▻ Looking for the best model to forecast the attendance

▻ Creation of an adaptative confidence interval

▻ Application to the quantification of the impact of the 2018 SNCF strike on a local network in Paris' region

▻ One proceeding admitted in a french conference

▻ Coding in Python

Junior Project Manager - Graduation Internship

Statistician - Internship

▻ Modelling consumers' loyalty to retailer

▻ Ranking the sections influencing the loyalty