/data-science-test

Primary LanguageJupyter Notebook

Data science technical-test

Repository content

  • DS_technical_test_tutorial.ipynb
  • Data: 4 datasets (description in DS_technical_test_tutorial.ipynb)

Test guidelines

The purpose of this test is to build a model (or several) in order to classify housing assistance requests thanks to the given datasets. This is a multi-class classification task, the metric to optimize and the datasets description are explained in DS_technical_test_tutorial.ipynb Jupyter Notebook.

Expected outputs (Python scripts or jupyter notebooks):

  • Data preparation
  • Modelization (at least two approaches including one with a deep learning framework: Pytorch, tensorflow, keras...)
  • Results analysis (BONUS : model explainability)
  • BONUS : api production ready

You will be evaluate on:

  • Code clarity, efficiency and production ready
  • The ability to justify and explain your choices
  • The ability to explain the theory of the methods/algorithms used
  • The final output score on the metric to optimize (good score below 0.7)