Tagging people in unanswered questions

One of the small improvements for this forum I proposed in my introduction post is setting up a list of people willing to answer questions and their expertise.

The table can be found at https://docs.google.com/spreadsheets/d/1VRJsngn8pdIr-duFQZmAxlWX-suy1xKpXi1rU7U4AtU/edit?usp=sharing

The aim is that whenever you encounter a question that remains unanswered for more than a few days, you can find someone on the list and tag them in a response, e.g.:

I don’t have have the time/expertise to answer properly, but maybe @martinmodrak can?

There is however no obligation to answer, so feel free to reject, especially if you are low on time, tired, … Take care of yourselves! If you can, it is best to explicitly say something, e.g.:

Sorry, @martinmodrak, can’t help, low on time.

But it is OK to just ignore if you are busy. The person who tags somebody should preferrably bookmark the question and check back in a few days and possibly tag someone else. (once again, no pressure - most of you are volunteers, do what suits you best).

One possible downside is that this could get people whose expertise is in demand overwhelmed with requests for help. To protect against this, I’ve hacked a simple tracker for the number of posts everybody has made in the last week to appear in the spreadsheet. This is a rough measure of how busy people are. Every person on the list can set a preferred number of posts per week and we should preferably tag people whose “quota” is the most free.

Feel free to request modifications in the description of your expertise and/or quota via comments in the spreadsheet or request edit access to the sheet (will give to anybody on the forum). This should also let you temporarily sign off (e.g. for exam period) by setting your quota to 0.

For now, I’ve just taken everybody’s last week post count, rounded and turned that into their quota. I have signed up those who opted-in in the responses (@bbbales2, @lauren, @paul.buerkner, @avehtari, @maxbiostat, @stijn, @emiruz and @sakrejda) and will try to continually ask active people to add themselves. If you want to be on the list, reply here, write me a message or comment on the spreadsheet.

8 Likes

Thanks Martin! I think this idea is worth trying out and seeing how well it works, especially if we can manage the spreadsheet well enough to provide a helpful guideline who to tag.

IMO it’s a good idea, and I’m happy to help. Many of the other contributors here are far more knowledgeable than me; so if I should make a mistake please don’t feel awkward pointing it out and correcting me.

Likewise, (and it’s just a suggestion), but for less authoritative contributors such as myself, hearting responses to signify agreement may be useful feedback.

5 Likes

Thanks! Great that you included the tracking

1 Like

Fine by me, thanks for organizing! :)

1 Like

Thanks @martinmodrak. You can include me too.

1 Like

I’m in. (I just need to figure out how to add a filter for mentions and I’ll be set.)

1 Like

Great idea! You can tag me for any GPU/OpenCL related issue or questions as well as Cmdstan installation issues on Windows or Linux. Not sure I provide any value in other areas.

1 Like

Count me in!

1 Like

Thanks for signing up. You didn’t provide topics you would prefer, so I guessed as:
@jonah - general modelling, @syclik - general modelling, Stan math, and @imadmali HMM, general modelling. Is that OK?

1 Like

I can answer Python (general use), PyStan and CmdStanPy questions.

1 Like

That’s a great idea, Martin! I’m available for model selection, R packaging, generic/weird error debugging and problems with documentation.

1 Like

Feel free to ping me on language questions.

Do you know how I can get email alerts when so pinged but not for every post?

I’m so overwhelmed by the forums now that I pretty much only try to follow developers and algorithms at an every-few-days pace.

1 Like

That’s fine, and also anything with the R packages, model comparison stuff (if Aki is unavailable), Stan language (if Bob is unavailable), and really anything else if you really need someone.

2 Likes

@imadmali HMM, general modelling.

You can put me down for general modeling, rstanarm, beta regression, spatial models, and HMM.

1 Like

I don’t think that’s directly possible, but every-few-days is IMHO an OK frequency. We’re a free service, operating under best effort :-) (and we reply to most stuff really quickly).

Thanks!

Feel free to add me to the list. I think I’d say the keywords I might be able to help with would include: general modelling, hierarchical models, multivariate models, mixture models, and Gaussian processes

1 Like

Sorry in advance if I’m misunderstanding, but that sheet simply lists the topic titles and author names for all posts here; it’s not surfacing any data that can’t be seen here already. Am I being naive in not empathizing with your discomfort?

The sheet is built from public RSS feed for this forum (https://discourse.mc-stan.org/posts.rss, via another, intermediary sheet for stupid technical reasons) which I’ve considered public data. It will not keep more than 2000 latest posts so it is not permanent either. However, I can totally filter you out if you don’t want to appear there.

Personally I’ve found this site preferable to Cross Validated for the stuff I want to do, but I am happy that other people are active there.

On second thought, @ChrisChiasson is right, the public list IS slightly problematic (at the very least it is IMHO not GDPR-compliant), and I should have not made it this way, sorry. I think I can make it work without the full list being visible, give me a few days, please.

4 Likes