Improving Data Quality

My birding observations are generally skewed towards species that will tolerate a human being within 30 feet of them…

3 Likes

The fact that some species have more photos than others is partly inflated due to most users observing from cities, but also merely due to some species being more abundant than others (relative abundance). For that reason, I don’t think of highly abundant species like honeybees as a data bias. To improve data quality we should strive for data accuracy. This means determining an optimal level of caution when IDing, and to be open to revise any ID if later shown to be incorrect. It also helps to caveat if any ID is less than certain, to prevent others from confirming it, although on the other hand their IDs are supposed to represent independent assessments.

All of the data including the common species are informative/meaningful when understood correctly, which although may require noting or correcting for any known sampling biases. Any species/group is informative to observe or ID, although it may be most informative to observe or ID a diversity of species. Re: most observations being from cities, it can also help to occasionally seek out more natural or remote locations to observe in for that reason, if possible.

3 Likes

@billryerson and @thomasgreen Welcome to the Forum - there are always interesting things to discuss!
All data are biased in some way.
I suspect, as well as the above mentioned comments, that the inability to see and identify is also a barrier. I could look for nematodes in my garden, but identifying them beyond Phylum would be beyond me. Similarly, I could search for tiny insects, but again, do not have a microscope to identify them.
There is also climate to consider - where I live, the only mobile organisms are birds and mammals in the winter. Everything else has migrated or is in some sort of diapause. In the spring I could set up emergence cones (outside of flood prone areas) but I don’t know how well they would work, or if they would be vandalized.
However, I do commend everyone who has developed a specialty, and apologise for not spending more time with the observations (I’m a Noctuid moth person). At least we are all doing something!

2 Likes

I wouldn’t worry too much about this. I think that most people using iNat data are going to fall into two (very rough) camps.

Group 1 will just be using tons of data from GBIF to answer big questions. They won’t spend the effort to dig details out of descriptions or tags or anything.

Group 2 will be focusing on a specific question or taxon. They will have a strong knowledge of this specific question or group and might find these aspects of data that you’ve proposed taking valuable, but predicting if someone will be interested in a certain question is very difficult. I would guess that data-wise, it would be much more valuable in general to just have a wide variety of lots of different observations with the standard data (time, location, etc.) than trying to add value via specific extra data to fewer observations.

That said, if you’re interested in “in-depth” observations, you could design your own project or try to find someone working on some question you are interested in and see if you can provide that kind of data. That way you know that your extra effort will go towards a project/question where it will be used/make a difference.

3 Likes

Lovely paper! I did a quick check of my own observations and I fall into Group 1 overall and Group 4 for insects.


This is my favorite bias graph from the paper. The urban influence is strong!

6 Likes

I agree this is as good as or better than a long descriptive note about associated species. If I see an interesting Sphagnum observation, I look for what else has been observed nearby to get a better feel for the habitat.

4 Likes

That’s interesting that according to the first graph you either love birds and nothing else or you don’t really observe birds at all.

1 Like

This. Let’s remember that iNat is not the sole repository. If comprehensive data is the point, there are more appropriate platforms.

Because field guides were essentially the precursors to iNat – they are at root a means of encouraging laypersons to engage with nature. Peterson, Audubon, Sibley: three well-known names in bird guides. Who is a well-known name in slime mold guides? Publishers are in business; as such, they need to see a market for a proposed book. Birds, shells, and wildflowers have more mass appeal than slime molds, flies, and flatworms. Vertebrates, on the whole, have more mass appeal than invertebrates, on the whole.

A serious researcher is going to use much more technical identification tools than a mass-market field guide.

4 Likes

Actually iNat helps filling the blanks about common species, herbariums are filled with hard-to-id weeds, and even though there’s a lot we can do to help with rare or not flashy species, we don’t need to forget that common species were/are often overlooked, even top 5 observed have big gaps on the map.

4 Likes

Great idea! Thanks so much!

1 Like

Hi All-

Thanks so much for all these awesome ideas/suggestions!! I’ve been in Vermont, documenting and learning about as much of the region’s astounding biodiversity as I can with kids in tow. I’ll be thinking about all these ideas.

Warm regards,

Raphie

2 Likes

@thomasgreen I love that you are aware and actually thinking about all of variables tho!

2 Likes

I love fungi, and have joined FunDiS as a volunteer to help with the search to find more of the rare fungi on the West Coast of North America…

If you are interested in rare things, fungi in particular, check out our West Coast Rare Fungi Challenge in both iNat and the website… Hopefully we can start to fill these knowledge gaps!

https://www.inaturalist.org/projects/fundis-west-coast-rare-fungi-challenge-ef75a1f8-d488-48f4-ae6c-d0a8c64a9fb3

https://fundis.org/protect/take-action

I totally agree with the observation bias for flashy, bright, easy to ID things, which is why I’m strongly driven to document weird forms or growth stages of common species, or as many (for example) black grey squirrels as possible, or buttercups with an unusual number of petals, or just every single mallard I run into lol. I try to mix it up between rare, mutants, and super common species. Also grass, moss and LIVERWORTS. Omg. Noone knows them and they are literally everywhere and there’s SO MANY DIFFERENT SPECIES.

3 Likes

Boy, I wish I had the patience to learn grasses (and sedges and rushes), mosses, and liverworts. Do I have all the books? Yes. I’m just so greedy to go new places, even just locally, and see what’s there that I’m not slowing down enough to look at whether a moss leaf is inrolled or not, or whatever distinguishes Cyperus from Scirpus, or even just find a liverwort that’s not our very common Bazzania or Conocephalum.

I should take a class or two.

3 Likes

One thing I try to do is, if an organism is common on the trail, not post my first encounter of it. I’ll post an observation from the middle of the trail, to spread things out a bit more. Not sure it helps in any scientific way, but it might make for a nicer map. ;-)

4 Likes

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.