datatogether / roadmap

Coordinating technical work & roadmapping additional services
8 stars 3 forks source link

Think of relationships to visualize in the superset instance #6

Open titaniumbones opened 7 years ago

titaniumbones commented 7 years ago

@sonalranjit has put up a superset dashboard for Arhivers at http://35.185.73.54:8088/superset/dashboard/archivers/. New visualizations are built with simple SQL queries on an attached datasource. RIght now it's pretty but it's not clear what kindso f visualizations would be most useful for us. Let's use this issue to suggest some visualizations. If you implement one maybe also tell us here!

titaniumbones commented 7 years ago

@mr0grog not sure if you will see this, but pinging quickly to say that Sonal could also set an instance up for web-monitoring if we wanted it. It's not clear what viz we would want there either, but since there's so much data, maybe we could find a way to look at it...

mhucka commented 7 years ago

This is nice to see. Good job, @sonalranjit!

mhucka commented 7 years ago

If I have comments/questions about this superset dashboard, should I post them as separate github issues in this repo, or as comments to this particular issue ticket?

sonalranjit commented 7 years ago

Totally up to you @mhucka , I would prefer separate issues tickets as long as you assign them to me since i started up the superset instance.

patcon commented 7 years ago

I've set up a Heroku-hosted superset instance, and it's got a github repo, if we wanted to take specifics over there? https://github.com/datatogether/datatogether-superset

It's kinda boring at the moment (no data sources), but when @sonalranjit gets some time, I think he was going to port the config from the previous self-hosted one. Done.

Once that's done, maybe we could clarify a way that folks can get limited access to the superset instance, so that they can play with data?

(Not yet sure whether email addresses can be obfuscated for certain roles, or things like that, so I'm not yet sure how liberal we should be with access. Any thoughts on this are welcome!)

patcon commented 7 years ago

@b5 What if we did an inventory of data sources? That's not to say we'd want to add all of them to this instance, but simply that they might be available for a later dashboard.