iNaturalist 2019 Team Retreat Follow-Up

Hi Everyone,

The iNaturalist team gathers for a yearly retreat to discuss our priorities for the coming year, and we met in a very wet and stormy Bolinas, California last week for this purpose. We also worked on a jigsaw puzzle and caught up on some reading during downtime.

Here is a un-ordered (except for the first item), non-exhaustive list of what we plan to focus on over the next year. Please keep in mind that some things may take longer than expected or might just not get done this year, other issues and priorities might pop-up, and that any big changes likely wonā€™t occur until after the City Nature Challenge, which ends on May 5th, 2019.

Looking it over, a major theme is ā€œimproving core functionality,ā€ making what iNat already does work better and more efficiently without adding a ton of big new features. The idea is to really strengthen what iNaturalist does right now so that we can add new functionality in the future.

  • Revamped notifications system and dashboard - This will have to be a revision from the ground up so it will take some time, but we acknowledge that this is part of the site which needs a big update. Iā€™ve started a topic in General where you all can share any thoughts you have about notifications. Note that new Feature Requests about notifications wonā€™t be approved, please share your thoughts in the General topic.
  • Implementing an embedded translation feature for comments and descriptions, similar to whatā€™s found on Twitter and Facebook.
  • Create a tool to help users split and merge their observations
  • Releasing a significant update to Seek in the next few months.
  • Rewriting the Explore/Observations Search page and adding more filters.
  • Adding a user-to-user trust system. For example, I could trust kueda and he would be able to see true locations of my observations.
  • Investigate ways to capture comments and ID remarks that are useful for making identifications and including them on the taxon page and in identification tools.
  • Bring up the iOS app to feature parity with the Android app.
  • Re-design/re-organize the Account Settings page. Itā€™s pretty messy at the moment and if more settings are desired, it would be best to make that page more manageable.
  • Continue to improve the Computer Vision model and the use of geospatial data in suggestions.
  • Continue to upgrade infrastructure and make the site more efficient behind the scenes as the community scales.

Those are the broad end-user related updates weā€™re looking to tackle over the next year or so. And of course weā€™ll be working to grow and improve the community (especially identifiers) as always, with some new members of the international iNaturalist Network on the horizon.

And there will be smaller improvements, bug fixes, etc. throughout the year. Weā€™ll be going through Feature Requests as much as we can.

We do want to thank everyone in the iNat community for all of your support and contributions, iNat would really be nothing without its users giving their time and energy to make it a positive place.

ā€” Tony

The iNat team (from left to right; photo by Abhas): Abhas Misraraj, Ken-ichi Ueda, Carrie Seltzer, Amanda Bullington, Patrick Leary, Alex Shepard, Tony Iwane, Scott Loarie

38 Likes

Awesome! This sounds like a great direction to go following the explosive growth of last year (and possibly this coming spring). The notifications update will be really helpful for us high volume users. The trust system may allow for more cases where people can use collection projects rather than the older project types. (though i think the older project type remains important for some uses). In-site translation would be awesome. I donā€™t merge and split stuff much other than copying observations but i know a lot of people want that feature. I am looking forward to seeing what you did with Seek! I donā€™t use it much since iā€™d just use regular iNat instead, but i think itā€™s a great option for kids, etc. Too bad our 3 year old is still a little too young but it wonā€™t be longā€¦ Really excited about the capturing of ID features too! And improvement of the computer vision too. Heck all of it. I hope you consider including captive species (or plants at least) in the algorithm since new users who rely a lot on the algorithm tend to observe those a lot.

6 Likes

If I want to split an observation I make a copy and I am always happy that the pictures are copied too. After that a remove from each observation the obsolete photos. What should this ā€˜split observations toolā€™ do?

1 Like

Investigate ways to capture comments and ID remarks that are useful for making identifications and including them on the taxon page and in identification tools.

Nice. Iā€™m looking forward to this. There are a lot of useful ID tips flowing through comments and it would be fantastic to be able to capture this knowledge on taxon pages (or something like that). Weā€™ll probably need to be able to connect single comments to several taxon pages as theyā€™re often about how to tell one taxon apart from another.

8 Likes

Itā€™s for cases where the user - generally a newer user - has multiple photos of different species in one observation and separate observations should be made out of each photo.

5 Likes

This will certainly be useful for paper wasps. Though on the other side, would there be a feasible way to get certain tips removed from a taxon page (or otherwise vote a ā€œtipā€ to be non-factual)? There have been a number of cases where rather incorrect notes had previously been used, and it really would be great not to have to see these pop up here.

6 Likes

What I have found most useful is when someone includes in their comments a link to a journal post that goes into great detail how to differentiate between species. Here is a recent one getting us straight on Phyla species in Texas: ([https://www.inaturalist.org/journal/lisa281/20234-turkey-tangle-frogfruit)

4 Likes

4 posts were split to a new topic: Train the vision model on cultivated plants

Thank you for all the great work you are doing to maintain and evolving the platform!

It would be great the get the possibility to choose the language of the mobile app within the app, or choose the language in which common names are displayed from within the apps settings menu.

3 Likes

This would be really great. I often ask folks for specific help so I can learn for the next time. It would be easy to check some little box to say that there is ID guidance within the observation for others to find. Or build it some other way. Iā€™d love to have this resource.

3 Likes

You have hit all my ā€œwantā€ buttons:) Such a good plan to strengthen the core before adding feathers, if I may mix my metaphors. Thanks.

3 Likes

Would it be possible that splits/merges could also be ā€˜suggestedā€™ by another user (similar to the ID process?) - e.g:

  • User X (potentially new) uploads 9 photos in one observation - 6 are a duck, 2 are a wasp, 1 is a dog
  • User Y looks at this, sees the issue, and selects the photos that should be split into two new observations
  • User X gets a notification, and a card appears on the original observation in the timeline - ā€˜Y thinks that this should be split into multipleā€™ (with the actual stepped changes detailed in the body of the card), and an ā€˜Agreeā€™ or ā€˜Disagreeā€™ button
    • Click ā€˜agreeā€™, two new obs are created and new user avoids a deep dive into iNat mechanics on first go (and User Y avoids having to try and explain where to click and what to do via the comments section)
    • Click ā€˜disagreeā€™ and User X can explain with a comment that weā€™re not looking at ducks, wasps, dogs, but actually the grass that theyā€™re standing on or whatever

(plus similar for merge)

18 Likes

Weā€™ll definitely consider that approach but I canā€™t make any promises. Cool idea, though!

8 Likes

very cool idea!

Great - more filters would be welcome! I have two specific requests:

First, would be to be able to filter for observations with missing metadata - i.e. are Casual because whilst they do have a
photo, they are missing observation date or location. There are a lot of records like this and they often come from users of the Android app, so I think theyā€™ve appeared as a result of syncing problems. If the user is still active and able to add the missing data, these are ā€˜quick winsā€™ for adding more observations to an undersampled area, as we have been trying to do in the Naturemap project (https://www.inaturalist.org/projects/naturemap-plants). At the moment I usually find them more by luck than judgement (i.e. by looking at records of a specific user).

Second, it would be brilliant to be able to filter for records with a placeholder, so as to rescue the often valuable placeholder text. It can indicate what in the photo a novice user has observed, or it can include a species name (with typo or not yet present on the system). Even better would be to have a mechanism that automatically adds the placeholder text to description or comment on upload, and of course to implement this for older records where the placeholder is still extant. But thatā€™s beyond a filter request :)

For our purposes, we would be able to search for these things within a collection project - whether thatā€™s through a form or via a url string is not so important.

Thanks for considering the request!

4 Likes

If the Explore search pageā€™s ā€œFront Pageā€ could end up being something other than a full world map load with dozens of observations loading all at once and whatnot, that would potentially be a nice usability boost for those of us on slow connections/connections that have a limited monthly budget of high-speed data. The google search page comes to mind- just a clean lightweight interface where you make your inputs first, rather than receiving a dump of data youā€™re not looking for.

13 Likes

Any news on this? Iā€™m just thinking about whether a tutorial on splitting and merging would be useful or if youā€™re close to having something ready.

7 Likes

Thereā€™s a lot of discussion of assisting users in splitting observations in this thread: https://forum.inaturalist.org/t/easy-way-to-mark-multiple-species-observations/278

No indication yet as to where this stands on the iNat teamā€™s roadmap for development.

5 Likes

Itā€™s not something weā€™ve gotten to yet.

2 Likes

Would really like to be able to merge observations. Have been taking multiple photos of the same organism from different angles hoping to help train the machine learning when I get a species-level ID from at least one of the photos I change all to that taxon ID but not good for anyone needing quantitative data or for people checking my IDs. Looks like hitting EDIT on the iPhone app allows adding photos to a single observation but didnā€™t catch/understand that till now. Must be many other novices in the same boatā€¦

2 Likes