datadrivenecologicalsynthesis / citizenscienceonlockdown Goto Github PK

You mentioned considering when cities went under lock down, are you also including re-opening/loosening of restrictions and how they differ across provinces right now?

I am testing out the issue feature. Is this how it works?

After today's meeting, I went into the Large GBIF dataset that we downloaded using Shuang's awesome key. There is a variable named: coordinateUncertaintyInMeters. From what I see, it has quite a wide range (NA, and from 1 to 4226661 [the Earth basically]). I went and checked the one observation that had that huge variability (here: https://www.inaturalist.org/observations/12281213) and it does show a huge area of uncertainty.

I think that we should add in a filter to keep only observations that have a reasonable accuracy given that we are selecting them based on their presence inside the city polygons.

• What do you want to pick as a reasonable precision?
• Should we exclude all the ones that have NAs for that variable? From what I can gather, most of the data has a pretty good accuracy (<200 m).

*I think inaturalist makes more sense for your project since its more for "everyday" users. GBIF and eBird is more "serious" birders

If your cities are on migratory routes that could be interesting like southern Ontario

Quick suggestion of a measure: iNaturalist has three grades: unverifiable, casual, and research-grade -- comparing the frequency of the three might be interesting

GBIF has eBird data but with a delay of a few years

Script for calculating park sums does not work for multi polygon features. Returns negative values.

Maybe grouping birds by landscape types that they are frequent in and see if their number of observations increased for more urban associated birds vs. woodland or grassland sp.

In BC, we have always been able to go outside whenever we want, but provincial parks were closed for most of March and April