This looks like a cool project -- always glad to see new tools for statistical collaboration and reproducible research.

How does this compare to what data.world [1] is doing? They recently released a Python SDK [2] as well.

[1] https://data.world/ [2] https://github.com/datadotworld/data.world-py