What would you like to learn about getting data from the system?

[pisum]

@jcook

For example, if you want to describe some use cases I could put together some Jupyter notebooks

muir earlier specifically mentioned Python as one of the languages he was trying to learn, and i think showing how to create something like the observation + taxon + observer count line graph in muir’s second example use case:

@muir

the state of Alaska (Alaska iNat passes 100k observations & 5k unique taxa)

… could cover many bases:

• the observation counts could be retrieved from /observations/histogram (which i suspect is probably the most useful endpoint for the majority of people who want to make time series charts)
• the taxon and observer counts could be retrieved by iterating through each month to get the total_results value from /observations/species_counts and /observations/observers, respectively.
  • you could show that such a request with per_page=0 is a way to get just the record count
  • you could also put a delay between requests just to reinforce iNat’s recommendation to try to limit request rates to about 1/sec.
• connecting it to a visualization package would be good, too.

something like sbushes’s export would cover additional bases:

• hitting /observations is probably another very common activity
• highlights PyiNaturalist’s ability (i think) to simplify multi-page requests and formatting / parsing of all the different kinds of values in the results.
• CSV exports (in cases where the standard download won’t work for some reason) seem to be a commonly requested thing.

@jcook

I think it would be worthwhile to show an example with Jupyter + Altair, and a similar example in Observable that accomplishes the same thing, in order to highlight the differences between the two platforms.

sounds reasonable. the Observable notebooks i highlighted as possible examples above do use Vega-Lite for some charts, and if we wanted the Jupyter and Observable examples for this effort to be exactly the same, i think it would be relatively straightforward to make a new Observable example that matches your Jupyter example.

@jcook

One minor correction: the “latest” docs you linked are for the pre-release version (dev branch); the “stable” docs are for the current stable release (master branch) and are better starting point, unless there’s a specific unreleased feature you want to test out.

i’m not sure which link you’re referencing. i did make a link to the Github Repo for PyiNaturalist, and the link there to docs seems to take you to the docs for the stable version. let me know if you still see an issue, or feel free to make a change to the wiki-style draft outline above.