Develop a system that effectively matches bikes with potential buyers to increase the likelihood of a sale and reduce the probability of item deletion. Find optimal synergistic strategies to in-cooperate boosted listings. Use this system in the ranking and the landing page. Also test if adding this as a threshold into the recommendations gives better results.
We setup the conda development environment with the following command.
make setup
Install requirements
make install
-
make lint -
make format
when creating an docker image the data is downloaded and prepared. Build test and production stages and runs tests.
docker compose build
Run app.
docker compose up app
src/driver.py defines the SQL queries, categorical and numerical features as well as the prefilter_features. config/config.ini holdes DB credentials.
REST API