To share or not to share private/obscured coordinates in a scientific publication?

I help manage the Never Home Alone indoor organisms project and we are gearing up to submit a data publication containing 20,000 records from the project. A number of the observations submitted to the project are private/obscured such that most users can’t see the precise coordinates for the observation. However, as a project manager, I am able to see the “true” location and accuracy of observations when I export the data.

My question is, should I hide this “true” location data in our published dataset, which will be made available (in a giant CSV) upon publication? I wasn’t sure if there was a standard way of doing this or what iNaturalist’s data use agreement is with private/obscured data. There are, essentially, two columns in the data, a “public location” column and a “private location” column. My understanding is the data use agreement allows project managers to see private locations. But I wasn’t sure if that allowance extends to project managers sharing that location data.

Since much of this data is from people’s homes, which is a sort of sensitive location, my instinct is not to share private/obscured location data even if it is allowed. Highly precise location data isn’t very necessary/useful for our project anyway. I just wasn’t sure if there was a standard way of going about this.

Thanks for any insight!

8 Likes

I don’t know the “rules”, but my thought is if the observer obscured the location that should be respected. Especially because it is peoples’ domiciles. I move my observations off my property enough, and well within proper limits, to serve both my privacy and iNat guidelines.

11 Likes

Might be good to just include the non-private coordinates and mention that the private ones are going to be a couple miles off on average. For most range based data I don’t think being a tiny bit off would really substantially change a species’ range.

Personally, I’d probably unobscure about half my coordinates if asked. I just don’t know what potions people make these days so I just over-obscure out of an excess of caution.

3 Likes

If my obscured locations were unobscured in a public spreadsheet, I would feel like my privacy had been infringed upon. Locations are obscured publicly for a reason, after all.

13 Likes

That makes sense. The reason I ask is just that the way iNat deals with obscured data is a little unique, since users are already making the choice to share private coordinates with project managers. I wasn’t sure if there was an explicit policy for what project managers are expected to do with the data when publishing research based on that data.

Since it doesn’t really matter for my project I’ll only share the obscured data. I suppose I’m mostly just posing the question for posterity, in case there are other projects where precise coordinates are important.

2 Likes

Are the locations connected to the usernames? If it is just an anonymous dot on a map indicating somebody has found an organism in his home, then I think it doesn’t matter. But if it shows user xy has found an organism in his home, that’s a different story.

3 Likes

No. If it shows my home it is violating the intention of obscuring. My name has nothing to do with it. It’s about the dot where I live on the map.

6 Likes

It doesn’t show “your” home, it shows “a” home.

1 Like

And trespassers are not welcome. That’s why I move them off instead of obscuring.

2 Likes

I suggest jittering all location data. That’s what I do when I produce maps/publications for work (the private lands I work on are protected by a privacy agreement). And include a disclaimer like “all locations jittered to obscure exact coordinates and protect sensitive information”.

6 Likes

Observations are obscured for various reasons. Some observations are obscured automatically (by iNat), which has nothing to do with the user who submitted the data. The location of an orchid is obscured to protect the orchid, not the user.

4 Likes

I don’t think that there’s a hard and fast rule for this, but I wouldn’t share the locations in this situation for several reasons.

  1. For observations with taxon geoprivacy, there’s likely a good reason for obscuring related to conservation.

  2. For observations that users chose to obscure themselves, they had a reason for doing so. Given the nature of the project, many of these are likely because it was the user’s house.

Going along with this, while users did choose to trust the project managers, it wasn’t explicitly stated that the data might be published when they signed up for the project (that I know of). They might have made a different decision to share if they had known this. While assuming that shared data would be published is something a scientist would do, many members of the public would not assume this.

I think that publishing the data would alienate some users or make them less likely to trust projects in the future.

  1. In this specific case, I don’t see compelling reasons why the unobscured coordinates are really needed for the readers to evaluate the work (though of course I haven’t read the paper…).

So, under what conditions might it be ok to share/publish coordinates like this? A few ideas:

  1. The project explicitly states that any shared locations might be published up front in the description. Users can make an informed choice to share/trust or not.

  2. The project manager/s explicitly ask individual users for permission to share their data (probably only feasible for smaller projects or datasets).

  3. As others have mentioned, the data is “anonymized” in some way. However, I don’t think publishing the data without a username would be sufficient. There are many ways someone would be able to figure out what coordinates go to what observation if they really wanted to, as the observations with usernames are still on iNat.

  4. There’s some scientific value to making the precise/unobscured coordinates public.

  5. There’s no conservation risk (ie, the taxa for which localities are published aren’t at risk of poaching, etc.).

9 Likes

When observers choose to allow the project manager to see the obscured coordinates, they should get a warning that there is no restriction on the project manager publishing them elsewhere (if that is in fact the case).

3 Likes

Maybe it’s just me, but if I tell someone a secret - and they know it’s a secret - I would expect them to honour that, not go and publish it to the world. I can see arguments for why this might be a different situation, but that’s my initial response.

5 Likes

I can’t wait to read your latest results!

3 Likes

Write a journal post on the project warning members - if you would share their location. I think many would have second thoughts about putting their obs in the project in the first place.

2 Likes

Don’t share the true coordinates of private or sensitive locations without permission.

2 Likes

Precisely, users have agreed to share the date with you, but not with the public at large. I absolutely don’t think private observation coordinates should be shared. Users trust that their contributions with be used in the way described, and researchers should respect that trust.

Separately, your research sounds fascinating! Thank you for sharing!

4 Likes

@cthawley provides an excellent answer. You probably should keep obscured datapoints obscured across all levels of use and publications. For any new projects, you can add disclaimers about data publishing to be more transparent to users but I don’t recommend amending projects already active and churning out data.

Additional tips:
You can justify to the journal/data sharing policies by stating that certain data were obscured due to various reasons, such as rare taxa or personally identifying data, and iNat policies, etc… I don’t think the journal would push back against something like this…and it won’t be the first time that iNat/citizen-science data are used in a study, i.e., the journal likely has dealt with similar scenarios.

If you publish raw data (e.g., a csv file) you can include the coordinates/precision from data that were public but, as cthawley points out, omit them from obscured points. Any reader/future researcher who really wants any given datapoint can see other metadata such as the iNat observation number/hyperlink in which they can then reach out the the original observer for permission. To that end, you can also state such in your “data availability statement” to the journal. I’ve had to do similar things in some of my publications and haven’t had much pushback as long as you can justify defending data privacy in your statement.

Good luck!

4 Likes

Depending on how common the species is, it could be very easy to match a species, date and obscured location to a username.

1 Like