Before I started using iNaturalist, I shared my nature observations via the Observation.org platform, as this is very popular where I come from.
However, for several years now, I have only been using iNaturalist, as I find it superior to Observation.org
Since I value iNaturalist not only as a platform for sharing observations but also as my personal nature journal, I find it very unfortunate that several years of my observations are stored on Observation.org and not bundled on iNaturalist.
I would therefore like to ask if any of you have experience importing observations from Observation.org to iNaturlist. Observation.org offers the option to export observations as csv, xlsx, SQlite, or KML files. However, I have never been able to import my observations into iNaturlist using these downloads.
Does iNaturalist even allow larger data sets to be imported into its own observation list? Or would you generally discourage the import of observations in order to avoid unnecessary duplicates in GBIF?
I would try and avoid creating duplicates at GBIF - unnecessary burden on data users and processors.
If you intend on having some observations (past and future) occasionally shared on both sites, you could adopt a default âcopyrightedâ license for your iNat observations (the only way to prevent their exporting to GBIF).
If simply repatriating past obs-org data to iNat then switching fully to iNat, you could set the default âcopyrightâ policy while importing data, then change it back to a permissive license (without applying it retroactively) for all your future iNat observations.
Alternatively, simply move all your data, i.e. delete your observations at obs-org and quit using that platform.
As for how to move data between platforms⌠no idea sorry.
if you forgo images and sounds and import just taxon, date, location, and maybe a description with a link back to the original Observation.org observation, that should be easy to do using iNatâs CSV import. these observations will be loaded as as casual grade observations because they lack images and photos, but that means that they wonât get loaded to GBIF and also that no one will try to identify them.
if you want to load the images, too, thatâs possible via the API, but i would tend to steer folks away from doing that under normal circumstances.
Do you have all the photos backed up elsewhere? That would make things much easier, as the Observation.org export facility isnât going to help you much. Unfortunately, the SQlite database doesnât include blobs for the photos, and the only one of the formats that includes links to the photos appears to be the KML.
If you need both the photos and the observation data, it would probably be wise to start by explaining your situation to the site admins and ask whether they can provide a custom export. The relevant FAQ includes a contact form for exactly this purpose, so itâs worth giving that a try before you attempt anything else.
Otherwise, obtaining all the data is far from straightforward. The KML file is an xml document, and the links are embedded in a CDATA section - but most of the other observation data youâll need isnât included. So you would first need to extract the photo links and the observation identifier from the KML, and then merge them with, say, the CSV export data. If you need to download the actual photos as well, it might take quite a while if you have several years worth of observations, as the site may well throttle any attempt at mass scraping of the image data.
Supposing you manage to restore the complete dataset to your computer, thereâs no reason why you shouldnât upload it to iNaturalist. If youâre worried about duplication, itâs easy enough to change the data sharing and/or licencing options on Observation.org - or you could simply delete the account entirely. At the end of the day, you own the data, so itâs up to you to decide how you want to share it.
If youâre determined to merge the full dataset (including photos) into your iNaturalist account, this is certainly doable - but thereâs no simple one-shot solution for this, so you would have to roll your own using the API. This will require some basic coding skills, but it shouldât be too difficult to achieve using e.g. pyinaturalist (assuming youâre familiar with Python).
Iâve managed to import my observations from observado.org with an Excel-file I made which can be used to translate from Observation.org-export to INat-ready import. Itâs far from optimal, but still the best I could relatively easy manage.
Pictures/audio are NOT imported
At import, the original observation-URL can be linked within the INat observation comment field. It would be better if this link to the original observation could be CSV-imported to an INat observation field (A field named âObservado URLâ exists allready in INat) , but adding a field value is currently not possible with the INat CSV-importer.
Information about the precences/abcence of photoâs in the original observation can be added in the INat comment field and the label field (makes bulk-editing on INat more easy).
Fill in the left table in EXACTLY the same format as those of the three example observations.
Copy your personal information in the correct blue column (TIP: leave the example observations so you can check if you fill in the right format
Be aware (prepare these things in another .xlsx-file):
The column order is different in the OBS-csv than the INAT csv. Shifting a column order can be easily done in Excel by selecting the column head, pressing SHIFT and dragging the column to the wanted position;
The notation of numbers can be different (e.g. â,â in Dutch becomes â.â in English);
The separator in GPS-locations should be with a â.â for INAT (is a â,â in the OBS-export)
If you want to obscure observations: you can fin in âobscuredâ or âprivateâ in the correct column. Leave empty if not needed;
There is a column for one or more labels (separate multiple labels by commaâs LABEL1,LABEL2)
Date should be transfered this format: â2015-05-13 11:01AMâ (or â2015-05-13â if you have no time)
Make use of Excelfunctions while preparing your import (like text string join)
Remove your âunsureâ observations (This doesnât excist in INAT) or change them to an ID-level youâre sure about
Taxonomy-matching can be usefull if you have al lot of observations and donât want thousands of import errors: https://www.gbif.org/tools/species-lookup (Check for the exact matches, edit the not-exact ones manually).
BUT most important of all: do this very very carefull, start very small with only a few observations. It takes a lot of time to correct things in INat (also with the Bulk observation editor). As long as you add no pictures, RG will be never reached and thus no duplicates in GBIF.
Wow! Thank you very much for all the suggestions, they are extremely helpful.
I didnât think that exporting would be so challenging.
Many of the photos relating to the observations only exist on Observation.org, so Iâll try playing around with the Excel/CSV import a bit.
Thanks @bazwal for pointing out that I am free to decide how and where I share my data. I hadnât thought about it that way before.
Big thanks @streepvaren for the detailed instructions that is super helpful!