MVP proposal for reddit post

jordansread commented 7 years ago

We are so close, but I think we need a few things to push this over the edge:

[ ] a processor that figures out which times the sites exceeded flood stage, and add a class for them (like f-4 f-5 f-6 f-7) if a gage was at or over flood stage for timesteps 4, 5, 6, and 7. js would set these to a different color for those timesteps. At each timestep I assume jquery would select all "gages" and set them to the normal color/class, and then select the flood class for that timestep and set that differently. Would need to make sure we also handled hover events on gages during a timestep transition.
[ ] add a second sparkline for "over flood" that is a different color and has no opacity until it is revealed. This would sit on top of the existing sparkline for the site.
[ ] use css transitions and transforms to do two things for the storm location: 1: make it move smoothly between various locations (transitions) and 2) make it so there is only one actual dot that we move around, avoiding the dual dot bug that happens on some mobile devices when you scroll or try to pinch zoom during the animation

--- should we also do the following? ------

[ ] use precip grid instead of the counties. Despite this being denser, I think it would actually improve the animation rendering and reduce the file size because the square/rect elements would be to much simpler than counties. Can draw all squares and use the states border as the clipPath
[x] use cumulative precip instead of hourly. It may be too boring w/o this (see #118 for this option)
[ ] have the storm change color/name when it is no longer a hurricane

wdwatkins commented 7 years ago

I like the change when no longer a hurricane

wdwatkins commented 7 years ago

Don't want to use the projections at all?

jordansread commented 7 years ago

Well, I think that ship has sailed, so no, not anymore, but they were useful for filtering sites.

jordansread commented 7 years ago

merged #146 but this is a little trickier, since the flood T/F is based on data of a different resolution than we use for the sparklines (hourly). I am pulling out a function to create the hourly data first before the classification.

wdwatkins commented 7 years ago

Yeah, forgot I needed another filtering step in there.

DOI-USGS / hurricane-irma

MVP proposal for reddit post #116