Time Series about review HOT 10 CLOSED

Ah, there was some miscommunication, i was only ever thinking of one aggregated profile. Definitely not one for every point.

from review.

This would actually be super useful to me in the very near-term. Do you have any idea how you'd like to go about this?

Like are we thinking we just replace the chart plot with a time series, or is this going to be a new page?

from review.

I would actually like the time series in a third element underneath the map and the chart, similar to those in DrIP. So, i was imagining it being stretched across the full extent of the page, but we could also just put it under the map and have space for a fourth element under the chart. It should only show if the profiles are available, as in we'd have to build an option to take in the HDF5s or NetCDF version of them.

from review.

Yea, I can do look for H5's. We can make the plot the full extent of the page for now, and change to be smaller if we ever need a fourth element in there.

What should it show by default? I imaging it can show a time series profile when you click on a point and/or select a few points, but we can hardly show all 56k curves at once. Should we just set an upper limit and say, no more than 100 curves at a time? Not too sure about this, super curious to hear what you think.

from review.

I suppose it can be an average profile of the selection.... then if you click a single point you'd still get the profile, but for 56k points, it would still be just one curve.

The downside is that you wouldn't be able to compare profiles with your selection, which might be something people would want to do

from review.

I was thinking we could pre-calculate diurnal profiles at each point to start. That's just 24 entries per point and we could even aggregate them by selection. I don't see why we couldn't compare profiles.

from review.

Oh compare profiles between two points, not two supply curve tables. Yeah, that will take some doing. In DrIP i have two maps and the option to turn of "location syncing". In it, the default is to share the selection between the maps, but if you want to compare two sites in one dataset, you choose the same in both maps, turn off syncing, and click points in both maps.

from review.

well, 24 * 56k points is still a very cluttered plot. Even 1 annual average point with 56K locations would probably be information overload, don't you think? We may have to start with some sort of aggregation no matter what.

I suppose if you don't think it will be that bad, we can test it out and see what it looks like. It might be just fine!

from review.

Gotcha gotcha. That is not as useful to me in the short term, so this may have to be a long-term goal overall

from review.

Implemented an initial set of elements for representative profiles. Now, if you toss these .h5s into the review_datasets folder for a project instead of the .csvs, reView will recognize them and render a timeseries veiwing portal across the bottom of the page.

from review.