Mark Silverberg

@rhyolight @chrismetcalf still, looks good and something I've been meaning to do. @socrata has a wealth of info. going to try2 hack on this

@rhyolight okay, you're right, a WIP and so a bit difficult for a newbie to get involved. Will watch from afar for a bit first. Great idea.

@skram Thanks, planning to refine over time. Big problem is that one Socrata resource can contain many streams of temporal data.

@rhyolight can you explain in another way? for example a crime dataset might have burglaries and motor vehicle violations?

@skram Exactly, or a water quality resource might have series for several locations, so each timestamp has several records.

@rhyolight got it. it's definitely going to differ by data publisher. IMHO, it may not be worth effort to ID by machine and instead [1/2]

@rhyolight do what you have started which is find the needles in the haystack and then annotate them by hand, to prove the concept [2/2]

@skram See https://github.com/rhyolight/soda-tap/issues/8 … for a specific example. Hoping to make this generic enough to be applied to all resources. [1/2]