This is not a specific feature request, but rather a request for information: I would like to know if iNat is likely to have a web uploader, or a 3rd party app, that will make it faster and easier to decide which photos to upload, crop them, group them into observations, and finally hit go, all in one place.
I’ve been using Wildlife Insights a lot for my work recently, and in most ways it is substantially less advanced and less convenient than iNaturalist. The contrast has gotten me daydreaming (during some startlingly long load times) about how a community science uploader could work.
Despite iNat being so advanced in so many way, the number of observations I upload is still limited not by my supply of photos, but by the time it takes to process and upload them. I have unused photos that could be many thousands of additional observations if I had time to process them. Please note that this is not a complaint, I am deeply impressed by the work the iNat team does, and grateful. I’d just like to hear please what prospects there are for a web uploader that makes the workflow simpler and faster for those prone to taking photos for several hundred observations in a day. I understand that the team has a long list of other priorities, and I’m curious if this is likely to ever happen.
That there are existing programs that can do the photo stuff like cropping, culling, and labeling photos, and they do it much better than software iNaturalist could make. Since we only have a limited number of developers and developer time, our focus is on the iNat end and not the photo processing end.
I’m curious as to where the bottlenecks are in your workflow. I currently use DXO Photolab to cull, label, and export my photos, but but there are options that are better for Digital Asset Management (DAM), like Adobe Photoshop Lightroom, Photo Mechanic, Capture One, and even the Apple Photos app is pretty good for that too (I don’t know any native Windows options). After geotagging my photos, I go through them and quickly mark and then delete all the ones that are bad. I then crop and export the photos I’m going to post to iNat. When I have a batch of 20-40 I put them in the uploader, add IDs and notes, combine if needed, and then submit.
i’m sure it would be possible to do something like this, but i think to achieve this all in a web-based solution that would be compatible with most browsers and devices and OSes might be complicated to cobble together from well-supported free code.
if you move the job of resizing images to the client rather than the server, i think that effectively means that you also have to shift all the metadata processing to the client, too. but then that code would also need to be able to read and save the data back to a particular data structure, and your image processing code would have to work with that same structure.
if you have the client do the image processing, then you also change the workflow from initiating upload upon selection to upload after some sort of review+edit process. i’m not 100% sure, but i suspect this would shift the burden of keeping a temporary copy of all those selected but not yet uploaded photos to the browser, which might reduce the number of photos a given device might be able to handle in one batch.
it would probably be much easier and more efficient to code a solution that would work for one particular client setup, but then only people with that setup would be able to use that particular solution.
It’s a good question. Like you, my uploading is limited by the time it takes to upload, and my photo backlog is now immense.
Rather than attempting to catch up with the available tools, I’ve dialed back my upoading and been quietly chipping away at a more automated workflow using R and darktable. That takes my field observations (mostly spoken audio notes) and uses those to automatically tag my photos (my editing Darktable’s sidecar files) for upload to iNat. It’s almost there.
The two steps I don’t see a way of automating in the iNat Uploader are whether or not an observation is captive/cultivated, and which photos should be grouped together into one observation. To my understanding, both of those steps have to be done manually on the iNat website during upload. (It would be handy if there were photo tags for these that the Uploader recognised.)
That all makes sense, thank you. My workflow is similar to yours, using the native Photos app on a mac. The thing that make it take so much time is that there are so many steps, and I have to come back to each photo multiple time. It gets imported into Photos, it gets cropped, it gets exported (which maintains the metadata), it gets dragged in small batches into the web uploader, it gets dragged onto other photos to make a joint observation, it gets IDed, it gets checked to make sure the metadata are all there and still correct after all that handling, and finally it gets posted.
I completely understand that 3rd party photo apps can do many of the things that need to get done better than anything the iNat staff have time to build, I just wish there was an app that integrated with iNat well enough to combine or bypass some of these steps.
My coding skills go no further than building some methods of data analysis in R, so I’m prepared to sound like an idiot, but I imagine that if the web uploader was built to accept all the require information from a 3rd party app, including the types of information @jon_sullivan is mentioning, someone would pretty quickly be offering a tool that made uploading much faster.
Thanks @jwidness I’ve not played around with that tool by @glmory. I really should, although I’m not nearly as comfortable in python as I am in R.
@dlevitis note that you can do everything prior to the uploader except those two things I mentioned. In case you haven’t discovered this yet, it includes species, description, location, annotations, and observation fields. The uploader gets all of that from properly tagged photos. It’s super helpful, as a lot of this is faster to add in a Digital Asset Management app, like DXO Photolab, Lightroom and darktable, than in the browser.