We are interested in decentralized data science infrastructure and networks. For example, imagine Pachyderm built on top of dataproject or IPFS or Beaker browser. We are specifically thinking about the protocols required to minimally share something like a data-puller (scraper, API script) and the data that is already pulled.

The space is growing, and it is not easy to see a clear winner here. We are looking for minimal approaches just to kick something along, and also to other people and groups thinking about this. I'm aware of qri.io.