A script is provided to simplify building the whole multi-module project.
Simply run
./prepare-docker-compose.sh
and be sure to put train.csv and test.csv in the data/ subfolder.
Done that, you can run any service with
docker-compose up db [service]
where service can be any of [dbloader, spark, or server].
You can find the report for this project as an Overleaf document over here.