A credit scoring model

Through a wide variety of mobile applications

  • Client Nusrat Nill
  • Date 25 June 2021

Project

The interactive dashboard meeting the specifications mentioned above and the score prediction API, each deployed on the cloud.
A folder on a code versioning tool containing:

  • The modeling code (from preprocessing to prediction)
  • The code generating the dashboard
  • The code allowing the deployment of the model as an API
    A methodological note describing:
  • The model training methodology (maximum of 2 pages)
  • The business cost function, the optimization algorithm, and the evaluation metric (maximum of 1 page)
  • The global and local interpretability of the model (maximum of 1 page)
  • The limitations and possible improvements (maximum of 1 page)
    A presentation support for the defense, detailing the work carried out.