mattparrilla / lake-brite Goto Github PK

First "slice" in each GIF frame is side closest to window. Fix data so that it works as expected

Different for lake and USHCN.

• lake needs no data
• USHCN needs empty slices

Robin suggested an amoeba-like representation of the lake data. Here are a couple of different takes on that:

1. No shape
   • don't clip the data to the lake, just use the entire display
2. No shape from middle
   • don't clip
   • put baseline plane in middle of display
   • values expand above and below
3. Abstract, amoeba shape
   • there is a baseline in the middle of the display, not a plane, but a rod like shape
   • measurements cause the shape to expand radially
   • make lake shape much more abstract, (maybe just a rod, truly amoeba like, but maybe right + left arms, like below)

Going with no.2 or no.3, consider coloring by value, not by level.

Lake is currently mirrored

Given a series of 2D frames, "pull" the frames through the "front" so that:

• the 1st frame of the 2D GIF is the front slice of the first 3D frame, the 2nd frame is the second slice of the first 3D frame, etc.
• the 2nd frame of the 3D GIF will have the 2nd frame as its first slice
• the 3rd frame of the 3D GIF will have the 3rd frame as its first slice
• etc

Where entire state of LakeBrite is a sprite in a frame of one single GIF

Should be some way to indicate where the animation is within a current loop. Could be a straight line, could be a line that moves around the entire outside that clears on repeat, etc

To avoid seizures

(0, 0, 0) should be bottom corner closest to the theatre entrance

no more generate_lake_brite_gif('Temperature') in .py files

Concept: Take the 15 lake monitoring stations and assign them to an (x, z) position (see Programming the Display to understand coordinate space). Then, pipe in the metrics, which will be displayed on the y-axis.

a number of things can be done:

1. project GIF onto single slice
2. project GIF onto all slices by staggering neighboring GIF pixels onto neighboring display slices (this will probably produce an interesting effect
3. consider transforming GIF in other ways

because of this

• Download Data
• Munge into 1D Array
• Function to convert 1D array into 2D array
• Consume 2D arrays (or images) in LakeBrite (pull through one frame at a time)

Roughly one week per strand, time pulls through the X-Y plane with X axis as entire year, Y axis is temperature (also encoded in color) and Z axis is the historical record.

At end of each loop provide an entire set of empty slices so you can tell where you are in the sequence.

Consider a diverging color palette with freezing as the splitting point. Throw out the more neutral, middle colors and go for the more intense blues and reds.

Turn Lake Brite into a 3D radar viewer.

Could pick some interesting weather, put it on a loop with decent speed.

(I modified the cron so that I'm no longer deleting the NOAA source data. Look into AWS to see if they offer historical data. If so, restore the original cron.)

From 50x15x10 to 25x15x20 due to the stagger

2015-12-07T00:35:20.905081+00:00 app[web.1]: Generating 3D Lake GIFs of Total Phosphorus
2015-12-07T00:35:24.535720+00:00 app[web.1]: Rotating matrix
2015-12-07T00:35:24.543958+00:00 app[web.1]: Increasing dimensions
2015-12-07T00:35:25.605849+00:00 app[web.1]: Normalizing values
2015-12-07T00:35:25.601789+00:00 app[web.1]: Stackining frames
2015-12-07T00:35:26.058644+00:00 app[web.1]: Mapping values to colors
2015-12-07T00:35:50.899373+00:00 heroku[router]: at=error code=H12 desc="Request timeout" method=POST path="/lake-gif" host=lake-brite.herokuapp.com request_id=51bfbbeb-7326-4b6c-8eac-63d511c7e55b fwd="162.247.89.226" dyno=web.1 connect=1ms service=30001ms status=503 bytes=0

https://devcenter.heroku.com/articles/limits#http-timeouts