Github Diff Viewer

Idea created by Bob Gill on Jun 15, 2018
    Open for voting
    • heheimann
    • tg_jbaetens
    • Bob Gill
    • fmastrodonato
    • sebastiangurt
    • joanponsa

    Most BOARD installations use proprietary datasets that should not be shared for a variety of reasons. There also exist numerous, public, free, useful datasets that could facilitate more detailed analysis. Now that we have a community site to connect, the next step is to engage. Github is a popular tool used for storing data repositories for a number of different projects. Originally, it was for code only, but does support development of projects where the files are binary. Using Github, we could collaborate on projects where data is freely available for public use. As we work collaboratively, the whole community benefits from being able to access a BOARD model. Here are some samples of datasets that could be collaboratively developed into BOARD models.

    • World Bank Macroeconomic Data
    • Weather
    • Social Media
    • Financial Instrument History
      • Stocks
      • Bonds
      • Derivatives
    • Government
      • Financial Data
      • Geographic Data


    For this to happen though, we need a way of reviewing submissions and comparing versions of the solution. For example, 3D modeling files are not readable by themselves. To really understand the differences from one version to another, a diff viewer is necessary. Below is a link to that kind of 3D diff viewer for use with github.


    3D File Diffs | The GitHub Blog  


    If the community is interested in collaborating on open-source style projects, a diff viewer compatible with github would be a very helpful tool to manage co-development across the globe. This tool would be able to compare the differences between one version of a model and another. The transporter tool already does this, but not in a way we can connect to github. I would like to see something like the transporter tool available for github.


    Maybe I'm the only data geek interested in working on public datasets, but I thought I would pose the idea for review.


    All comments welcome.