pyvideo-data status: March 16th, 2016

Note: This is an old post in a blog with a lot of posts. The world has changed, technologies have changed, and I've changed. It's likely this is out of date and not representative. Let me know if you think this is something that needs updating.

What is pyvideo-data?

pyvideo-data is a collection of Python-related conference video data in JSON format.

This is an update of what's going on with pyvideo.org and pyvideo-data.

Status

In January, 2016, I wrote up a blog post about how I was ending pyvideo and why and roughly how I felt about everything. Many people wrote emails to me about how they used pyvideo and how pyvideo going away would affect them. Some people were interested in taking over the project.

Some of those emails started discussions with Naomi and others at the PSF which got me thinking about the importance of the data that we had collected and the importance of continuing to curate that data.

Sheila and I talked about this at length. It's hard to really get across just how ridiculously large a project pyvideo was and how much work and expertise is required to do it well and how awful we felt because we were underequipped for it. We knew we didn't want to reboot pyvideo, but we thought it was important that someone spent time bootstrapping pyvideo-data into a healthy community project. Everyone is busy, so we decided we'd start this process and see how far we got.

This blog post is pretty overdue, but we wanted a few key things to happen before we started the ball rolling.

  1. We wanted a mailing list so that everyone who is interested in curating, using or analyzing Python-related conference video data could collaborate.

    Courtesy of the PSF, we have a pyvideo-data mailing list. If you're interested in curating, using or analyzing this data, please join.

  2. We wanted the infrastructure to allow for drive-by editing of the data. Things like fixing names, fixing typos, adding additional related urls, etc--these things should be easy to do, easy to review and easy to merge.

    The repository on GitHub is the master copy of the data. Drive-by editing of that data is possible using GitHub's interface and the barrier to entry is having a GitHub account. If you don't have an account, you probably know someone who does. Once you make an edit, it creates a Pull Request and the changes you made are automatically validated. At that point, someone can review and merge it.

  3. We wanted the tooling and infrastructure to make it possible for people to add conference data.

    We wrote clive which has a validation subcommand to validate JSON data and make sure it's "correct". While the software and the validator probably have issues, this infrastructure is there and we can improve it over time. clive also has a fetch subcommand that pulls down metadata and generates the video JSON files for it. This needs work, too, but it's there and I used it to add DjangoCon 2013, DjangoCon 2014 and DjangoCon 2015.

These are things we talked about over the many years we worked on pyvideo, but we could never figure out how to do them with the architecture and resources we had. I claim this is one of the reasons pyvideo just didn't work out. These are all in place to some degree with pyvideo-data today.

So what now?

If you're interested in:

  1. curating the data (adding new conferences, editing existing data, improving the data, ...)

  2. using the data (static site generation, static site of specific slices of the data, ...)

  3. analyzing the data (conference speaker diversity, conference description complexity, how many people are talking about Django, ...)

please join the pyvideo-data mailing list.

The next set of things we need to work on are:

  • bootstrapping the community that will curate this data going forward

  • writing documentation for processes and tools so that it's easy to participate

  • solidifying the schema for the category and video data

  • improving the validator so that it catches more issues before changes are committed

  • improving the infrastructure and tools so they're easier to use

There are probably some others that I'm forgetting about.

So what about pyvideo.org?

There are no plans to kill the site, yet. I'm hoping that as we bootstrap pyvideo-data, a bunch of people who are interested in setting up a statically generated site based on the pyvideo-data data will self-organize into a team that works on that project.

Depending on how that works out and what they end up building, I'm game for transferring the pyvideo.org domain.

Until then, we'll leave the existing site up for now.

Want to comment? Send an email to willkg at bluesock dot org. Include the url for the blog entry in your comment so I have some context as to what you're talking about.