A set of Docker based scripts to visualize differences in large datasets. Visualizations are png files created by seaborn. Eventually, the goal of this tool is to provide at a glance visual QA of the GRLS UBC dataset.

We begin with two complete databases for comparison. Tables included in data-definition.json are dumped as tsv files. These are compared via datacompy, which creates summary csvs. The summary csvs are then massaged and visualized.

1. Make sure images are built for the given host (docker images). If not, build them via: docker build --target=mysql-env -t=<some name> . and docker build --target=python-env -t=<some name> . These containers based on these images execute all steps of the script.
2. Create a workdir directory with subdirectories: one, two, comparison and app. Define the workdir in init.sh.
3. Make bash scripts executeable and run ./init.sh