# PyMC Community Team Important, especially with PyMC 4.0 coming up and all the questions about how to port models. Important tasks: 1. Invite team members: Christian Luhmann, Dan Weitzenfeld, Michael Osthege, Bill Engels, Alex Andorra, Meenal, Oriol Abril, Ricardo Vieira, GMCobraz_T, Colin Caprani, Dawar Qureshi 2. Invite those that say yes to Slack & give them "highest level of trust (4)" on discourse 3. Create better labels for topics (type of model: linear regression, GP, ODE, installation problem, shape issue, convergence, general modeling question, technical question) 4. Actually apply labels to all posts, starting with most recent / unanswered. 80% issues are already answered elsewhere, requires follow-up. More inexperienced users can answer easy questions, do light formatting changes of the post, and flag more difficult questions in slack for questions they can't answer themselves. ## Agenda - PyMC Community Team 2022-01-06 6 Jan. 2022 - What is the community team? - Discourse - What else? - PyMCon? * Notes * Split community team need to do versus nice to do * Main motivation: Discourse had a lot of questions who were naturually answering so this may be one area we can focus on * Other open question: Whether pymcon can be in the purview of communinty events * Is PyMCon team separate * Permanent members, with stable group of people * Hardest part will be time commitments, answer discourse questions, reach out in organized way * Allocate responsibilities * Availability of discourse is patchy, sometimes theres not enough resources or attention to topics. * Lets have a more stable baseline, if something takes too many efforts, then free consulting or free effort * Initially when learning GP questions because its focused, also a plus sign of joining community team to learning. Assigning those question to specific volunteers * Better labeling, but not doing it systematically * Is it installation related issue or questions? Is it specific posterior predictive * Discourse should be everyone on team, rather than external team. People are responsible for discourse is efficient, so everyone else can make the mos * Right now every question is acknowledged, act as moderators, right away help out * GP or time series, then we can tag people that know about it, otherwise we can help out * Even if someone doesn't get answered, theres no info followups etc * We wont know all the answers, questions get a response, people they can be subscribed to notifications. Keep checking those labels. At least a lot of the labels are not updated. Maybe on discourse to keep things fresh * Mission of community team: Grow the community * is to answer questions on discourse ourselves * We've also been neglecting marketing, we do so much cool stuff * So don't have a dedicated way of telling people about that * Should we include marketing in community? * The success of packages often hinge on the success of a community * Short term we have discourse, getting that in order to boost efficient * So other people have cycles, longer term, there's lots of other things to be considering * Think big, even if in short term we get Discourse working the way we want * Cross pollinating, getting developer discussions moved wholesale and developer presence on discourse, seeing that work both ways. * Having developers see whats happening, installation related are core developers don't see those struggles * Blog posts that are tweeted about the kind of thing that is large as it actually is - StackOverflow? - PyMCon - Discourse - Help desk (current discourse activity) - Recurring events/activities? - More entries in the "Sharing" topic (e.g., blog/notebook posts/discussions/breakdowns?) - Q&As? - Paid Discourse account? - https://www.discourse.org/plugins - Team procedures - Delegation of discourse questions to team members? - Elevation of discourse questions to github issues? - "sub-optimal" questions (see [this pymc-examples issue](https://github.com/pymc-devs/pymc-examples/issues/258))