quanted / cyan_app

django app for cyan
0 stars 0 forks source link

machine learning predictive modeling #15

Closed aignatius closed 5 years ago

aignatius commented 7 years ago

As a developer, I would like to use GEE machine learning functions (ee.Classifier) to find relationships between weekly cyan bloom concentration and various influencing parameters (precip, temp) so that I can predict the likelihood cyan blooms will increase, decrease, or stay the same 1-4 weeks into the future.

Acceptance criteria: given a cyan timeseries (derived from geotif image collections), lake boundaries, and influencing parameters, I would expect a timeseries with future 1-4 week cyan concentration predictions.

time-dynamic datasets PAR Wind precip temp longwave radiation shortwave radiation

time-static datasets percent landcover in watershed size of watershed lake size lake bloom frequency lake avg annual max CI lake avg annual max bloom area