undefined | Better HN

0 pointsjg0r32mo ago0 comments

I work in a large conservation organization focused on rare plant conservation.

iNaturalist is sometimes used by our ecologists/biologists as a starting point for collating occurrence data.

The iNaturalist data itself is likely specifically being pulled from gbif. Then they go private/specialty databases that have more spatially and taxonomically accurate records.

But iNaturalist data is often not considered high quality enough to be publishable by itself (wide brush statement) in my field of plant conservation.

We've tried to have some conversations with iNaturalist and they weren't really interest in talking, gave me pause on what their motives as an organization are.

But conservation tools are few and far between, and iNaturalist is a really powerful tool for initial data exploration.

0 comments

MattBlissett2mo ago

GBIF track the use of data we provide to scientists, where they later publish papers citing that data [1]. For iNaturalist, the list of known citations is at [2]. In most cases a download of data from GBIF will include data from more than one dataset (iNaturalist is one of over 110,000). To find how particular records in a download were used — or even if they were discarded — requires reading the paper.

As an example from the list, "Aedes albopictus Is Rapidly Invading Its Climatic Niche in France: Wider Implications for Biting Nuisance and Arbovirus Control in Western Europe" [3] cites 5348 iNaturalist records.

[1] https://www.gbif.org/literature-tracking

[2] https://www.gbif.org/resource/search?contentType=literature&...

[3] Paper https://doi.org/10.1111/gcb.70414 citing GBIF data download https://doi.org/10.15468/dl.gzdq3f

PS we are recruiting an engineer, but the deadline for this position is today: https://gbif.link/senior-data-engineer

jg0r3OP2mo ago

Ah, what a wonderful role. I'm sad I've missed the application period, maybe one of the only jobs in tech I'm specifically qualified for :(.

I should add the asterisk that my anecdotes on the presumed legitimacy of publishing solely iNaturalist data comes from conversations I've had in the American endangered/threatened plant conservation community. More common or non-American species occurrence data from iNaturalist likely has more legitimacy being used directly in publications.

In the US we have a series of government/NGO controlled databases that house sensitive species data, so often our scientific community has to operate through them to get access to publishable information (of which the raw data is then often obscured, just used for analytics). In my experience iNaturalist data is often a good starting place for determining which government bodies/NGOs a biologist should start reaching out for requesting data access.

I love GBIF and have a priority this year of making sure that my organization plugs-in what we're willing to share via IPT or Biocase!

sq_2mo ago

> But iNaturalist data is often not considered high quality enough to be publishable by itself (wide brush statement) in my field of plant conservation.

As someone who recently started using iNaturalist, I've been curious about this. I think it's an awesome platform and really cool that people can share what they find, etc, but I noticed that people would pile on with species-level IDs on pictures that were obviously ambiguous between different species known to exist in the vicinity.

I of course want as much data as possible to be available to science, but it piqued my interest about whether a negative feedback loop of misidentifications to future identification models could form.

j / k navigate · click thread line to collapse