changing quality grade: research, needs ID, unverifiable

[Ken-ichi]

Hi folks,

In short: we're changing the way quality grade works for observations.
Instead of just "research" and "casual," we're

1) narrowing the scope of "research" to only include stuff ID'd at the
family level or lower,
2) introducing a "needs ID" state for observations that could be come
research grade but need more IDs, and
3) introducing an "unverifiable" state for observations that cannot
become research grade

We're also adding the ability to mark observations as "reviewed." This
will happen automatically when you add an ID, but you can also do it
explicitly for observations that you can't ID but don't want to see
again.

You can test these change at

http://gorilla.inaturalist.org
username: preview
password: 313phant

You can test out filtering by reviewed like this:

http://gorilla.inaturalist.org/observations?reviewed=false&quality_grade=needs_id

We'd love your feedback. The gorilla test site should allow you to
login and add IDs and stuff like you would on the live site, and it
will have no effect on the live site.


THE LONGER STORY

This all came out of some conversations here in the Google Group:

https://groups.google.com/forum/#!topic/inaturalist/F3HLBHQG0A0
https://groups.google.com/forum/#!topic/inaturalist/q4KiiVCwN-o

which we attempted to synthesize (after much haggling) into some new features:

https://github.com/inaturalist/inaturalist/issues/576
https://github.com/inaturalist/inaturalist/issues/611

There are a couple different needs that are being addressed here, most
of which are dealt with in detail in the above links, but the
overarching goal is to help people become better identifiers by
allowing them to focus on observations that need their attention.
We're also hoping these clearer delineations and filters will help us
build better tools focused on the experience of adding
identifications. We (the iNat team) are focusing on two main groups of
users at the moment: identifiers and new users. New users are
obviously important if we're going to achieve our primary goal of
connecting people to nature through technology, and our secondary goal
of creating enough data to be scientifically useful. We want to make
new naturalists, and we want them to make new observations. But in
order for them to get something out of iNat in the form of comments
and IDs, we need to expand and empower our core community of
identifiers. So this is a step in that direction, for the obvious
reasons of providing better filtering and the ability to ignore
things, but also because it helps clarify what we're all trying to
achieve here: helping people out by sharing our knowledge, and making
good data.

Anyway, looking forward to your feedback! We're hoping to get this on
the live site in the next week or two.

-ken-ichi

[Kent McFarland]

Worked for me nicely. I wasn't sure how something became unverifiable. Clearly those with no photo do. But is there a specific way to make it so? Or is that automatic under certain situations? Can I mark something as unverifiable in other words? Sorry if being dense. 

Kent

____________________________

Kent McFarland
Vermont Center for Ecostudies
PO Box 420 | Norwich, Vermont 05055
802.649.1431 x2

-ken-ichi

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Scott Loarie]

Hi Kent,

The attached diagram was useful to me:

1. if newly created obs have dates, locations, photos/sounds, and are wild they become 'Needs ID' otherwise they become 'Unverifiable'.

2. Needs ID is automatically turned off when the community ID is of rank species or lower (these obs would move into the Research bin)

3. Needs ID can also be voted off by the community (e.g. blurry photo)

4. Once Needs ID is turned off, obs with a community supported ID of rank family or lower become Research, obs with no community supported ID or with a community supported ID of rank higher than family become Unverifiable

5. Research and Unverifiable obs can always move back into Needs ID if the community votes Needs ID back on

Does that make sense?

--

--------------------------------------------------
Scott R. Loarie, Ph.D.
Co-director, iNaturalist.org
California Academy of Sciences
55 Music Concourse Dr
San Francisco, CA 94118
--------------------------------------------------

[Ken-ichi]

One addendum: you can vote on whether an observation should be in "needs ID" using the "Still needs ID?" vote under the Data Quality Assessment.

[Tim. Reichard]

I think the set of improvements from this update are going to be a nicely polished improvement on the ideas proposed a few months ago. I like the simplicity of the single ID status over the four separate flags (Research Grade, Needs ID, ID Please, No Finer ID Possible) that were previously proposed.

Tim

[Charlie Hohn]

a couple of assorted ideas:

it would be nice to be able to filter out ones own observations. Obviously I can't make my own observations research grade but when I search for stuff in Vermont to review i see those too. Maybe have ones own tagged as reviewed by them, since in a sense they are. But maybe it only bothers me because i have too many observations.

it would be fun to have an option to get these observations to display in random order, I think that would help find 'buried' stuff for ID.

On Thu, Aug 6, 2015 at 6:30 PM, 'Tim. Reichard' via iNaturalist <inatu...@googlegroups.com> wrote:

I think the set of improvements from this update are going to be a nicely polished improvement on the ideas proposed a few months ago. I like the simplicity of the single ID status over the four separate flags (Research Grade, Needs ID, ID Please, No Finer ID Possible) that were previously proposed.

Tim

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

============================
Charlie Hohn
Montpelier, Vermont

[Charlie Hohn]

or... maybe this is happening but there is no way to filter out reviewed observations yet? Or maybe i am just missing it? If so sorry.

[Ken-ichi]

Anything you ID should be marked as reviewed, so that includes your
own observations. I think I forgot to generate those for past
observations on gorilla, though, so I'll try and do that. Once the
data is there I think
http://gorilla.inaturalist.org/observations?reviewed=false&quality_grade=needs_id
should not show your observations (presuming you've ID'd them and
you're signed in).

On Thu, Aug 6, 2015 at 4:52 PM, Charlie Hohn

[Ken-ichi]

Actually I take that back, it seems to be working, and you can filter
by reviewed. Here are all the Vermont observations that need ID that
you haven't reviewed:

http://gorilla.inaturalist.org/observations?reviewed=false&quality_grade=needs_id&place_id=47

[Charlie Hohn]

Awesome

Sent from Gmail Mobile

[James Bailey]

I thought unverifiable might have a blaring red warning box with the text on it like RG and Needs ID! Too bad :)

[James Bailey]

I have a question, say I was with someone and we both saw but didn't photograph something. I put the observation, and he can agree with it because he saw it himself. Will that remove the "unverifiable" tag?

[Carrie Seltzer]

I love this! Great work. This makes it so, so, so much easier to find things that need ID confirmation at the species level. Turns out there ARE monarch butterfly observations on iNat that haven't been confirmed.

Suggestions:

1. Add UI for "reviewed" in search.  I see how you can use the url to exclude stuff you've already reviewed as mentioned to Charlie in this thread, but I think there should be a user interface option for that in search that allows you to select either stuff reviewed by you OR stuff not reviewed by you.

2. Fix copy. Looks like there's extra copy in the search field because there's a check box that reads as "w/ sounds needs ID" but should be "w/ sounds".

I can't wait until this is live on iNat!

Carrie

P.S. If anyone who uses google or facebook to log in to iNat has trouble logging in to gorilla, just make another account for testing using an email address (you can add "+anything" to a gmail account to make it accept an existing email address, e.g. yourgma...@gmail.com) for test purposes. That's what I have to do for gorilla because I log in with google.

[Kent McFarland]

Perfect explanation and very clear. Thanks for taking the time. This is real good.

The only thing I'd quibble about is the family level. I don't see much use for research grade at that course of ID. Genus and better would be my preference, but it isn't a big deal. Do we send research grade observations that course to GBIF? Seems like that would be just for numbers inflation for them or something. Anyhow, not a big deal to me. This is a great improvement. Thanks for the hard work on it by all. 

Kent

____________________________

Kent McFarland
Vermont Center for Ecostudies
PO Box 420 | Norwich, Vermont 05055
802.649.1431 x2

[Charlie Hohn]

Yeah, for plants family level doesn't seem so useful for research grade. Other taxa are different probably. 

Sent from Gmail Mobile

[Donald Hobern]

Genus or better would be an appropriate "research-grade" cut-off for most groups I know.  

Donald

[Donald Hobern]

Oh yes - I should have said.  I'm really pleased to see the new model as a whole - this is a wondoerful step forward.

Donald

[Scott Loarie]

Hi James,

It won't, Research Grade has always required some form of independently verifiable evidence (e.g. photo or video)

On Thu, Aug 6, 2015 at 6:36 PM, 'James Bailey' via iNaturalist <inatu...@googlegroups.com> wrote:

I have a question, say I was with someone and we both saw but didn't photograph something. I put the observation, and he can agree with it because he saw it himself. Will that remove the "unverifiable" tag?

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[krancmm]

A few questions, of course.

In Ken-Ichi's original post he made two statements.

 

"New users are obviously important if we're going to achieve our primary goal of connecting people to nature through technology"

Is there a mechanism that easily shows why an observation is unable to be verified?  I know bugguide struggles with how to tell a user that an ID is impossible because the image is too small/oof/doesn't have the correct views but without offending or discouraging the user.  In iNat, the same circumstances occur.  As a new user I'd find it off-putting if a bunch of my observations had an 'unverifiable' slapped on them without some gentle explanation or guidance on what is necessary for an ID.

It's also possible that a new user doesn't realize the potential of various projects.  If a clear image observation were added to the correct project, an ID might be possible.  Does one just suggest it be added, leaving it to the new user to do all the work, or can we automatically add it to an appropriate project? 

 

"..our secondary goal of creating enough data to be scientifically useful....introducing an "unverifiable" state for observations that cannot become research grade"

One question/concern here is the possibility that some well-meaning users will slap a coarser grade (say, from species to genus because they only know the genus).  That allows the observation to become research grade but is of less value scientifically.  I'd immediately "opt out" of community IDs if that happened with mine.

Linked to above paragraph, I have a number of insect observations that I see are "needs ID".  They are not "unverifiable" - the specialists that originally IDed them aren't on iNat and probably never will be.  They could easily become research grade, and scientifically useful, if my previous suggestion (buried in the discussion about converting Facebook users to iNat) was considered.  Very simply, a checkbox next to an observation shows that it has already been vetted by bugguide, BAMONA, known expert.  It automatically becomes research grade if it provides, either in description or separate field, the link to the vetted ID.

I'd also agree with others that family level is often, but not always, too coarse to be of much use for research.

Thanks for working on improving the iNat experience for us all.

Monica

[Scott Loarie]

In both the google group

https://groups.google.com/forum/#!topic/inaturalist/q4KiiVCwN-o

and the issue

https://github.com/inaturalist/inaturalist/issues/576

the RG cutoff was proposed at 'finer than family' not 'family or finer'.

But 'finer than family' still includes subfamily, supertribe, tribe, subtribe, supergenus. Is this what you mean or do you mean 'genus or finer'?

We'd be 'loosing' (e.g. moving from RG to Needs ID) 38k with 'family or finer'

http://www.inaturalist.org/observations?quality_grade=research&lrank=epifamily

70k obs at 'finer than family'

http://www.inaturalist.org/observations?quality_grade=research&lrank=family

and 75k with 'genus or finer'

http://www.inaturalist.org/observations?quality_grade=research&lrank=subtribe

Ken-ichi, did you implement it at 'family or finer' for a particular reason? I don't have particularly strong feelings about the cutoff.

Scott

[Scott Loarie]

"New users are obviously important if we're going to achieve our primary goal of connecting people to nature through technology"

Is there a mechanism that easily shows why an observation is unable to be verified?  I know bugguide struggles with how to tell a user that an ID is impossible because the image is too small/oof/doesn't have the correct views but without offending or discouraging the user.  In iNat, the same circumstances occur.  As a new user I'd find it off-putting if a bunch of my observations had an 'unverifiable' slapped on them without some gentle explanation or guidance on what is necessary for an ID.

Thats one of the reasons we settled on the word 'unverifiable' because its less of a value statement than something like 'incomplete'. The idea being that lots of people might never intend to add photos to their observations and we didn't want them to feel like they were doing something wrong. But yes when someone votes off Needs ID without providing a fine ID its probably polite to lead an explanitory comment. I try to do this when I vote no on 'does the location look accurate' and probably should also do it when I vote 'captive' so the observer knows why their data quality metric has changed.

 

It's also possible that a new user doesn't realize the potential of various projects.  If a clear image observation were added to the correct project, an ID might be possible.  Does one just suggest it be added, leaving it to the new user to do all the work, or can we automatically add it to an appropriate project?

Voting off Needs ID should only be done if you are really confident that no-one can provide a better ID. If you can't ID it but think someone else can, you should mark it as personally reviewed so it show up in your searches for things to ID. Like this one which 'I' couldn't ID but apparently gregpauly and xantusiahunter could http://www.inaturalist.org/observations/1633782

so maybe marking as reviewed and adding it to a relevant project is a good workflow

I'm not sure how often I will confident enough to vote off Needs ID because I'm always curious if someone else could ID something where I could not, but I guess a blurry spec of a photo or an obs of something I'm really familiar with where the photo is obscuring the key distinguishing character. But curious to see how this feature works. Remember you can always vote Needs ID back on if you think it is Identifiable 

 

 

"..our secondary goal of creating enough data to be scientifically useful....introducing an "unverifiable" state for observations that cannot become research grade"

One question/concern here is the possibility that some well-meaning users will slap a coarser grade (say, from species to genus because they only know the genus).  That allows the observation to become research grade but is of less value scientifically.  I'd immediately "opt out" of community IDs if that happened with mine.

Yes, I admit this is confusing but we should remind people in comments that adding a coarser ID after a finer ID implies a disagreement, clearly this is kind of confusing as it happens alot

 

Linked to above paragraph, I have a number of insect observations that I see are "needs ID".  They are not "unverifiable" - the specialists that originally IDed them aren't on iNat and probably never will be.  They could easily become research grade, and scientifically useful, if my previous suggestion (buried in the discussion about converting Facebook users to iNat) was considered.  Very simply, a checkbox next to an observation shows that it has already been vetted by bugguide, BAMONA, known expert.  It automatically becomes research grade if it provides, either in description or separate field, the link to the vetted ID.

This is something that might be handled by building the much discussed reputation system. But for the moment, iNat's quality grade metrics are all based on independent peer review within the iNat community, so maybe its important to add that context when we're talking about data quality. Because your right, Needs ID doesn't mean the quality is poor, just that it doesn't have independent peer review within the iNat community

 

I'd also agree with others that family level is often, but not always, too coarse to be of much use for research. 

Thanks for working on improving the iNat experience for us all.

Monica

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Ken-ichi]

Carrie: our plan is to completely revamp the
http://www.inaturalist.org/observations/id_please as an
identifier-centric interface that will focus on quickly identifying
observations or marking them as reviewed and moving on. The "reviewed"
functionality is pretty much only of interest to identifiers so we
probably will not include it in the general search UI. I'll look into
the copy fix.

Regarding what ranks get research grade: the way I implemented it is
that anything at family or finer can get research grade. Happy to
change it to finer than family though. IMO the rank cuttoff is pretty
arbitrary since taxonomy is subjective. Also, keep in mind that the
situation is slightly more complicated than I originally described.
Scott's diagram illustrates it accurately: observations will remain
"needs ID" until they are ID'd to species or lower. The only way they
can be "research grade" and ID'd at family or lower is if they get
voted out of "needs ID."

[James Bailey]

There are certain groups especially fungi, mildew, etc. which never make it near family without serious laboratory work, I guess these will never receive research grade with the new system?

[Sam Kieschnick]

Big change here!

Overall, I do like it though.  I think it helps build credibility in the "research grade" and also recognizes that some observations simply cannot be ID'ed further without more information (other photos/angles/ages of organisms, etc...).

Thanks!

[Ken-ichi]

For fun, I just added reviewed and captive buttons to
http://gorilla.inaturalist.org/observations/id_please, which should
give you a sense for what it would be like to work through a limited
set of observations (I just tried plowing through some from a local
park). Pagination is not the right approach here and those buttons are
a bit confusing so maybe not ready for prime time, but you can see
where we'd like to go in terms of functionality at least.

[krancmm]

Playing around and noticed something a tad odd:  when I agreed with an observation, it was also made into a "fave" - not my doing.  Don't know at what point it occurred as I was messing with turning on and off "still needs ID":  http://gorilla.inaturalist.org/observations/1806218

Whoops, it's happening with every observation:  http://gorilla.inaturalist.org/observations/1806219
Even after I deleted my ID, the "fave" still remained.

Assume it's a little fixable glitch?

Monica

[Sam Kieschnick]

I'm catching the same glitch -- when I click "yes" to "Still needs ID," it adds my name to the "fave..."

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/rxohZR7QQ_k/unsubscribe.
To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Jon Sullivan]

Hi all,

I'm just catching up with all this thread now, here and on Github. There's a lot that's been said!

In general, I'm all for it. It's a really useful step forward. With one exception.

I find it odd that my photographed plant observations identified by many people and tagged as cultivated are displayed as "Quality grade: Unverifiable". This doesn't do much to feed the enthusiasm of users tracking changes in flowering phenology, or mapping out parent source populations of invasive plants, or figuring out what the flowering resources are for pollinators in an urban area, etc. In many urban and rural habitats, these plants are a real part of their ecosystems, even if it was a person rather than a bird that planted their seed.

I get why these non-wild observations aren't fed up to GBIF, but it would be helpful if the identifier tools that will be built on top of these new quality grades were still accessible for observations of non-wild organisms. I see that Scott separated "captive/cultivated" from "unverifiable" in one of the earlier flow charts on Github (https://github.com/inaturalist/inaturalist/issues/576). I've attached a revised version of that for what I would see as being ideal.

While we may be right to be wary of a flood of kittens and rose cultivars swamping iNaturalist, I see a clear place for observations of non-wild species in a naturalist's exploration of human dominated habitats. It would be a shame if the new system made that more difficult.

Cheers,

Jon

[Charlie Hohn]

Hmm.. I agree. They aren't unverifiable. And as Jon says there are lots of cases where those observations are useful/important. 

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[krancmm]

I completely agree with Jon and really like the revised flow chart...there's a huge value in knowing what cultivated plants might support a healthier ecosystem.

Monica

[Tim. Reichard]

Since the Still Needs ID? question will be an important part of entering IDs, could this question and its answer options be placed prominently next to every opportunity to enter an ID?

For example, in the right column of an Observations >> Needs ID page like
http://gorilla.inaturalist.org/observations/id_please?utf8=%E2%9C%93&q=&search_on=&quality_grade=needs_id&identifications=any&captive=&place_id=&swlat=&swlng=&nelat=&nelng=&taxon_name=Halysidota&taxon_id=120218&day=&month=&year=&order_by=observations.id&order=desc&rank=&hrank=&lrank=&site=&tdate=&d1=&d2=&filters_open=true&view=

And in the bottom-left of a single-obs. page like
http://gorilla.inaturalist.org/observations/1804316
On this page, it could also be retained where it is, in the collapsible Data Quality Assessment box -- it seems to belong there, too. But if that is collapsed, it would still appear prominently by the ID-entry box in the bottom left.

Tim

[Scott Loarie]

The reason we chose the term 'unverifiable' is that it doesn't carry value or judgement like 'incomplete' or 'not valuable', but rather that these observations aren't research grade or en route to becoming research grade. I agree that you argue that a someone's pet fish can be 'verifiably' identified as a Oscar or a Goldfish, but by that same logic you could also verify that an observation without a date or a location could be 'verifiably' identified. But here we're choosing for ease-of-use to lump observations that aren't Research Quality and don't Need IDs to become Research Quality (e.g. those without photos, or locations, or wild subjects) into a single category with a label.  So I guess our use of 'unverifiable' as that label is somewhat a narrower definition meaning an observation that because it lacks certain characteristics be they a photo, or a location, or a wild subject makes them less able to be assessed by the natural community as natural history observations. If anyone can think of a better label than 'unverifiable' that describes the observations lacking photos, dates, location, or wild subjects please propose one, but changing the structure to make a whole separate category for the small minority of captive observations on the site IMO isn't worth it.

Also, in defense of the label 'unverifiable' I think you can make a pretty good argument, just as you can with observations without dates or photos, that IDs of observations lacking wild subjects really aren't as 'verifiable' from a natural history standpoint. Location is a key character for identifying many species, with cultivated critters, we don't have access to relevant location information. I can ID fishes of California, but I have no idea how to ID some random fish some random kid might have in their California fishtank because I can't use location as a worthwhile character to aide my identification. Likewise with cultivated critters are often strange pseduo-domesticated varieties that don't exist in nature. This opens up a whole world of candidates for IDs that are really outside the realm of Natural History, e.g. 'this is a Pink Lady cultivar of some domesticated rose variety'

But again, all this aside, even if we did complicate things with special categories and emphasis for the tiny tiny minority (<2%) of observations on the site representing captive critters (and again we have to recognize that adding special categories and emphasis creates more buttons and settings and dials and knobs and paragraphs of documentation that will add even more complication and confusion to the site which has a real cost in terms of usability when I think most would agree that iNat being too hard to use is a major bottleneck towards growing the community). What would be the motivation? Would it be to encourage observations of cultivated species? I think here it pretty controversial on whether this is the kind of use of the site people would like to see more of. IMO, I don't really want to encourage captive observations and would rather build a community that emphasizes recording and sharing observations of wild things over one that emphasizes sharing observations of goldfish and tomatoes. Also, since the first thing people new to iNat observe is often their dog or their potted plant, we will always see a stream of observations of captive stuff, so it would be nice to work towards encouraging these new members to move beyond their dog and get outside and explore nature. I think by including these captive observations in alongside observations without locations, or without photos in the category of observations that aren't really the focus of the site is a subtle way to emphasize that while its fine to post observations without photos or locations or of captive stuff, the focus of the site, at least from a community standpoint,  is to record and share natural history observations with evidence.

Just my opinion and apologies in advance for the long email,

Scott

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

[Scott Loarie]

Realizing that few would read my previous long ranty email, I just want to emphasize my main points:

a) If you wanted to search for captive organisms for whatever your purpose is you can always do that.

b) But in terms of explaining and displaying the structure of how observations on the site are grouped/displayed/and otherwise emphasized to new users, moving from three categories: 'Unverifiable' (14%), 'Needs ID' (38%) and 'Research Grade' (48%) to four categories: 'Captive' (2%), 'Unverifiable' (12%), 'Needs ID' (38%) and 'ResearchGrade' (48%) to add special emphasis to this 2% of captive observations that aren't core to the mission of the site IMO adds a cost of complication and documentation that far outweighs the benefit of avoiding a slightly imprecise definition of the 'unverifiable' label by not lumping captive observations in with observations that don't have photos, locations, or dates in the 'Unverifiable' group.

[Charlie Hohn]

I think one of the issues here is the type of organism. Agreed someone's pet doesn't matter. A pansy planted in the ground in June that dies at first frost doesn't matter. A tamarisk planted in a yard next to a stream does matter though as does a planting of native plants that bees are eating. For that matter so does a planted crabapple tree that blooms six weeks earlier than before due to climate change or weird weather. 

Here in Vermont I set the line between things that overwinter and things that don't. In California it would be things that survive without irrigation. It's certainly a difficult line to draw and maybe not worthy of pursuing. But I also don't think we should simplify Inaturalist too much. Honestly, we don't need every user ever and there are better sites out there for users so casual they can't figure out how a cultivated tag works. Like Instagram. 

Just my biased opinion too. 

--
Sent from Gmail Mobile

[Scott Loarie]

The core issue here is not what the system allows you to do (you can post and search for pretty much anything on inat if you're inclined) - but what's emphasized/default by the user interface. I'm arguing this should be the most typical use case. The system on gorilla, by not including captive things in the needs id bin, does not surface captive obs to IDers by default. Remember if you wanted to ID captive stuff you could futz with the search filters to find and ID these captive obs.

So the core question is whether the default behavior should be to check an extra box to include captive obs in a search for things that need Id or to check a box to exclude captive obs from this search.

IMO, most ID work flows would prefer to exclude captive obs so the former would be a better default. 

I agree that people might post obs of captive stuff to iNat And hope for an ID, but we're seeing that the ID community on Inat does not tend to ID captive stuff, but are rather ignoring these obs (ie excluding them by default as the system on gorilla does), I suspect the reasons for this are twofold, a) they aren't interesting to IDers, and b) IDing captive stuff is much harder because you can't use range as a character to narrow the possible choices. these are my reasons for avoiding captive obs when I add IDs but would be great to hear perspectives from other IDers

Do you folks disagree and think on average IDers would like to see captive obs in their searches by default? Because remember you can search for anything on iNat, it's just a matter of checking  a box and updating the filters, but imo we should make the defaults the most typical use cases. 

Scott

Sent from my iPhone

[Charlie Hohn]

Oh, sorry, i twas just talking about whether they would be marked as unverifiable. I agree the main datastream for ID Please should exclude non wild/naturalized things especially considering the ability to add them via changing the filter.

============================
Charlie Hohn
Montpelier, Vermont

[krancmm]

I believe I'm suffering from too many over-firing synapses, but here are some random neural firings...

I think that the 2% "captive" for plants is very possibly an under representation.  Even in parks or wildlife refuges it's often difficult to know whether human intervention occurred at some point, so a user is likely to leave the field blank (assuming "wild").  If the location and taxon have a reasonable relationship, that observation would theoretically skew the data if the plant was indeed cultivated.

..."an observation that because it lacks certain characteristics be they a photo, or a location, or a wild subject makes them less able to be assessed by the natural community as natural history observations."  I understand iNat core goals and requirements differ from other sites, but eBird trucks along quite well in their research endeavors without the need of photos; many entomologists are very interested in assessing the relationship of an insect to a plant, wild or cultivated.  The validity of assessments depends on the people who use them - many possible roads to assessing data.

Every definition I've seen for natural history, and there are lots, emphasizes relationships.  While researchers have different goals in their studies, I'm having trouble understanding how my insect observations can be research grade when almost every one of them is associated with a cultivated plant.  If the core mission of iNat is "all wild", shouldn't only "wild" insects on "wild" plants be included in research grade?

The main value of seeing or IDing captive plants is, for me, the possibility of determining which plants have the best ability to increase the diversity of associated biota.  I'd also concur with Mr Hohn that there is value in the phenology of cultivated plants.

Given some of the feedback, including mine, it appears that "unverifiable" does carry a negative connotation (as it does in the thesaurus entries I checked).  How about just "NOT Research Grade"?

Monica



 

[Scott Loarie]

Hi Charlie,

Ok if you're not talking about including Captive observations in the Needs ID bin, than that change you're proposing is less controversial and ignore my previous argument. But keep in mind that the diagram would not be the one Jon pasted in but this one where Captive obs would not be in the Needs ID bin. Thats what you all mean, correct?

But so now the question would be whether to lump Captive (2%) of obs in with the obs with no photo/date/location observations (12%) into a single 'unverifiable' bin (14%) or whether to keep these separate.

The pros of keeping them separate you're bringing up seem to be:

a) its not technically accurate to refer to captive obs using the label 'unverifiable' (if you can find a better label than 'unverifiable' please propose one!)

b) the set of captive obs are of particular interest to the community to warrant special categories and emphasis in the UI and documentation on the site

The cons of keeping them separate seem to be:

a) the 2% of captive obs my not be of  interest to warrant first-order grouping like this. If captive gets special attention, than why not also 'obs with sound' as a special first order category?

c) things can get complex very quickly and again, people who want to search for these niche categories can always find them using the advanced filters. But we're trying to design easy to use UI/documentation/cartography and adding additional categories really does have a cost. So the idea was to start with the simpler option (3 categories: Unverifiable/NeedsID/Research) and then if that doesn't work well build more complexity (eg 4 categories: Captive/Unverifiable/NeedsID/Research) rather than start with potentially unnecessary complexity.

Does this sum it up or are there other pros/cons I'm forgetting? I'd still argue that the cons outweigh the pros, but curious to hear what others think.

Scott

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[James Bailey]

With plants there are two main ways to classify them.

Way 1: natives count, long established and self-sustaining non-natives count, naturalized, waifs, etc. do not.

Way 2: all plants count, including waifs, as long as they appear to be naturalized and not a random occurrence that hopped over a fence nearby.

It is often hard to draw the line but I've never had any confidence issues deciding if a plaint is obviously self-sustaining or not, or whether it was obviously from ornamental/cultivation nearby. It is personal I suppose how people decide what counts and what doesn't. 

James

[Charlie Hohn]

I always count waifs unless I think they definitely won't Persist long (ie things that aren't winter hardy). Because why not document them?

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
Sent from Gmail Mobile

[James Page]

I also completely agree. I'm getting into this conversation a bit late, but I was about to make a similar comment. A good example is here this observation, which as an invasive species would be really important to track:

http://gorilla.inaturalist.org/observations/1806132

James

[James Page]

I posted a comment on this before continuing reading through the thread, which wasn't the best approach.
Basically I agree that certain non-naturalized species should able to be research grade therefore can be used with more confidence by organizations like GBIF, invasive plant council, etc. While most of the species tracked by those organizations would in fact be naturalized, iNat users may not know this. The red eared slider observation here is a good example (http://gorilla.inaturalist.org/observations/1806132).

Maybe the term Captive is the problem - on one hand we're thinking people's gardens and pets, but the species I'm thinking of aren't captive, necessarily, they're escapees or introduced.
What if for the question: Is the organism wild/naturalized? we refer to something like, is the organism living/growing in the wild? This will allow for those observations that aren't pets, but are non-native.

On Monday, August 10, 2015 at 7:53:20 AM UTC-4, Charlie Hohn wrote:

I always count waifs unless I think they definitely won't Persist long (ie things that aren't winter hardy). Because why not document them?

On Monday, August 10, 2015, 'James Bailey' via iNaturalist <inatu...@googlegroups.com> wrote:

With plants there are two main ways to classify them.

Way 1: natives count, long established and self-sustaining non-natives count, naturalized, waifs, etc. do not.

Way 2: all plants count, including waifs, as long as they appear to be naturalized and not a random occurrence that hopped over a fence nearby.

It is often hard to draw the line but I've never had any confidence issues deciding if a plaint is obviously self-sustaining or not, or whether it was obviously from ornamental/cultivation nearby. It is personal I suppose how people decide what counts and what doesn't. 

James

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Charlie Hohn]

Hmm, I would have marked that turtle as naturalized unless it is living in an artificial pond and being fed. It may have started as a pet but has gone feral.  Maybe I'm wrong in that but... to me captive is an actual pet or livestock animal, not an escaped animal that survives on its own. With plants, anything that can survive on its own that wasn't planted (and can survive a Vermont winter if found here) I consider naturalized. 

So maybe I am using the concept wrong, but in these cases I would not have hit the captive tag.

On Tue, Aug 11, 2015 at 12:06 PM, James Page <page.j...@gmail.com> wrote:

I posted a comment on this before continuing reading through the thread, which wasn't the best approach.
Basically I agree that certain non-naturalized species should able to be research grade therefore can be used with more confidence by organizations like GBIF, invasive plant council, etc. While most of the species tracked by those organizations would in fact be naturalized, iNat users may not know this. The red eared slider observation here is a good example (http://gorilla.inaturalist.org/observations/1806132).

Maybe the term Captive is the problem - on one hand we're thinking people's gardens and pets, but the species I'm thinking of aren't captive, necessarily, they're escapees or introduced.
What if for the question: Is the organism wild/naturalized? we refer to something like, is the organism living/growing in the wild? This will allow for those observations that aren't pets, but are non-native.



On Monday, August 10, 2015 at 7:53:20 AM UTC-4, Charlie Hohn wrote:

I always count waifs unless I think they definitely won't Persist long (ie things that aren't winter hardy). Because why not document them?

On Monday, August 10, 2015, 'James Bailey' via iNaturalist <inatu...@googlegroups.com> wrote:

With plants there are two main ways to classify them.

Way 1: natives count, long established and self-sustaining non-natives count, naturalized, waifs, etc. do not.

Way 2: all plants count, including waifs, as long as they appear to be naturalized and not a random occurrence that hopped over a fence nearby.

It is often hard to draw the line but I've never had any confidence issues deciding if a plaint is obviously self-sustaining or not, or whether it was obviously from ornamental/cultivation nearby. It is personal I suppose how people decide what counts and what doesn't. 

James

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
Sent from Gmail Mobile

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

[Scott Loarie]

I agree with Charlie, red-eared slider are definitely naturalized here in California. 

But more broadly, 'Captive/planted' is meant to capture zoo animals, pets, potted plants, and tended garden plants. I know its not black and white, particularly with the latter category, but I wouldn't worry too much.  All we're talking about here is a handful of default UI labels and search components that give a first order impression of the kinds of observations that 'Need ID', are 'Research Quality' and are not that 'Unverifiable'. Its meant to categorize observations broadly, not cover every edge case and all the complicated nitty-gritty details and semantics. The advanced search filters and more subtle labeling has in the past and will continue to show you that. As I keep emphasizing, there's nothing that prevents YOU from posting observations of captive things, identifying observations of captive things, or exporting observations of captive things for your own purposes.

And after we try this, if the consensus is that these three categories don't display things as comprehensively as people want, then lets try splitting off a 'Captive' category from 'Unverifiable' or some other complication. But again, this is mainly to try to come up with a cleaner, easier to use handful of default UI labels, cartography, and search components to help identifiers new to the site. Currently we have 2 default categories - 'Casual' and 'Research', moving to 3 is already a big step up in complication. Lets wait and see how that goes until moving up to 4. And again, there's nothing before or after this change that prevents YOU from posting observations of captive things, identifying observations of captive things, or exporting observations of captive things for your own purposes.

Insert IMO at the beginning of every sentence above,

Scott

[Ken-ichi]

As far as I can tell, the only controversy here is over semantics:
some of you don't like that our definition of "unverifiable" includes
observations of captive organisms, but as Scott has repeatedly labored
to explain, "unverifiable" is really just our best attempt to pick a
value-neutral, concise label that means "cannot become research
grade." The only alternative label I've seen proposed here is "NOT
Research Grade," which won't work because it is both long and
inaccurate: "needs ID" observations are also "NOT Research Grade."

Categories and semantics are *always* subjective, and while we
definitely want to include as many of your opinions as possible when
defining them, we could argue over them forever. If no one can propose
a better label than "unverifiable," I think it would be more
constructive to focus the remaining critiques on *utility*: what are
the effects of this change in categorization?

1) Will these changes affect the way captive/cultivated status affects
the "research grade" label?

No. Captive / cultivated organisms have never been candidates for
research grade status.

2) Will these changes affect the data we share with data partners?

Yes, but not with regard to captive / cultivated status. We don't
share those observations with GBIF, ALA, or Calflora now, and we won't
after this change. We actually implemented captive / cultivated
filtering at the request of Calflora because they were alarmed at all
the cultivated plant observations they were getting from us. What *is*
changing is that we are going to stop sharing observations ID'd at the
family level or higher. If a data partner decides they *do* want
observations of captive stuff, we can accommodate them. We only
recommend that data partners consume research-grade observations, but
we can easily give them other kinds of observations.

3) Will people still be able to upload observations of captive /
cultivated organisms?

Yes.

4) Will people still be able to identify observations of captive /
cultivated organisms?

Yes.

5) Will identifiers still be able to find observations of captive /
cultivated organisms?

Yes.

6) Will people still be able to use iNat to record (or download) data
about cultivated plant phenology, interactions between wild and
non-wild organisms, or anything else you can currently record on iNat?

Yes.


#2 is the biggest change to how we are using these categories and it
has nothing to do with captive / cultivated status. For those of you
concerned about captive / cultivated observations not getting
identified, keep in mind that a) as Scott mentioned, many identifiers
already ignore these without any special filtering, b) you can still
find and identify these, and c) very few people self-report captive /
cultivated status anyway, so these will probably continue slipping
into the stream of observations that identifiers are paying attention
to.

-ken-ichi

[Charlie Hohn]

I'm OK to try the 'unverifiable' label for captive obs. Though I suppose I wasn't the most outspoken opponent of it. Sorry if I added to the wandering of this conversation into the realm of the potential value of these observations, regardless of what they are called.

[Ken-ichi]

Hey folks,

Made a few changes on gorilla:

* voting "yes" on "Still needs ID?" should not affect faves
* changed review button to a checkbox and moved it down to the
identifications form
* added reviewed UI to observation search filters
* made the RG cuttoff *finer* than family instead of family or lower

I think that deals with most of the bugs reported (thanks!), so I
think we will try and release this on Friday.

[Jon Sullivan]

OK. I can live with having to click another button in the filter to see the non-wild things that require IDs. I accept Scott's point that we shouldn't add complexity to iNat for just 2% of the observations.

Nevertheless, while it is semantics, I'm still uncomfortable labelling verifiable observations of non-wild things as "unverifiable". It's neither accurate nor elegant. And iNat is elegant!

Plus, as Monica mentioned, quality grade "Unverifiable" has some negative connotations in a way that "Casual" didn't.

The best solution I've come up with is to just make quality grade "unverifiable" =  "moderate" and "research" = "high" (I'm avoiding "low"). I imagine anybody who notices the quality grade on their observations would rather see their unverifiable or non-wild observation labelled as "quality grade moderate" rather than "quality grade unverifiable". And it makes sense that we only push the high grade observations to GBIF.

Using "Moderate" and "High" solves three things:

1) "Unverifiable" has somewhat negative connotations. "Moderate" doesn't.

2) Labelling verifiable non-wild observations as "unverifiable" is inaccurate. "Moderate" quality is accurate for non-wild observations because as a whole they're less useful.

3) You can do plenty of research on observations that don't have quality grade "research". "Moderate" quality observations are generally less useful for research than "High" quality ones.

 "Moderate" and "High" score highly on both accuracy and elegance in my book.

Would that work?

Cheers,

Jon

[Ken-ichi]

I hear you about the inaccuracy of the label. I don't particularly
want the categories to be so explicitly ordinal like that because I
feel like that makes even less value-neutral, but your post did remind
me: why don't we just use "casual" instead of "unverified"? I think we
were pursuing "unverified" b/c we were originally going to separate
quality grade and ID status, but since we're lumping them, I feel like
"casual" makes sense again. So the categories would be "Casual,"
"Needs ID," and "Research." "Casual" doesn't exactly mean "cannot
become research grade," but neither do "moderate" or "unverifiable,"
and it's pretty value-neutral.

[Kent McFarland]

I like casual

--

____________________________

Kent McFarland
Vermont Center for Ecostudies
PO Box 420 | Norwich, Vermont 05055
802.649.1431 x2

[James Bailey]

We use "casual" already and it works so I think we can stick with that over "unverifiable".

[Jon Sullivan]

Cool. Yes, I'd be happy with "Casual," "Needs ID," and "Research." "Casual" is more in keeping with the fun and friendly vibe of the site than "unverifiable".

[Carlos Galindo-Leal]

Hi guys,

For NaturaLista the translations would be

1) "grado de investigación"

2) "necesita identificación"

3) "no verificable"

In Mexico we provide grants for 16 taxonomists that are reviewing IDs. It would be great if those people could

have more weight in their IDs.  http://conabio.inaturalist.org/pages/ayudacuradores

Furthermore, it would be great if slowly, we could labelling "not possible to identify" or something else, those species than can not be identified by photo.

Saludos

Carlos

[Ken-ichi]

Ok, I'm going to go with "casual" then. Carlos, weighted
identifications is a whole separate topic related to reputation
systems. Regarding things that you cannot identify, we've kind of
encapsulated that here by allowing people to vote whether or not an
observation still needs an ID. The "Casual" grade basically means
things that cannot become "Research" grade, which includes
observations that have been voted out of needs ID because they cannot
be identified.

[James Bailey]

It just occurred to me that I'm no longer going to be able to keep track of my unidentified plants (that is, plants that I don't know the species of and previous had ID please as the tag).

I relied on searching with the filter of "ID please" to find my observations tagged with ID please, I assume this is no longer available as a feature?

I wonder if we could have some sort of second "fave" list, for personally keeping track of certain important observations. Could be a "pinned" list or something. 

[Ken-ichi]

So FYI, this is now live. The data quality grades are slowly being
updated for older obs, but newly created ones should have them set
correctly.

James, you can still see your "ID Please" observations at
http://www.inaturalist.org/observations/kueda?has[]=id_please, but
what I prefer to do is to use the rank filters, e.g.

http://www.inaturalist.org/observations/kueda?lrank=genus

You could also look at all your obs that still need ID

http://www.inaturalist.org/observations/kueda?quality_grade=needs_id

These links aren't going to work until the data get updated, which is
in progress.

[Charlie Hohn]

I agree that it would be nice to have a way of self-identifying ID uncertainty. We could keep the ID Please functionality in a sense and change the tag to 'Identifier Uncertain" or something but if there is concern about increasing site complexity maybe that isn't ideal.

Thanks so much for working on this! I can't wait to see it take effect.

C

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Charlie Hohn]

One other question... when looking at an observation page rather than the Needs ID page... I see I can check needs reviewed but then I don't quite see how to save it without making a comment. Will it autosave right when i hit the check box or do I need to do something different?

This looks wonderful and I anticipate using this Needs ID page a lot.

[BJ Stacey]

I created a field (helpful) that allows me to find these types of observations.  Ones where there are some good comments added that help with identifications in the future.  Maybe not exactly what you're looking for but perhaps something in the right direction.
http://www.inaturalist.org/observation_fields/2552

I do know that having it done as a field allows anyone to use it and not just me, but it was the best solutions that I've come up to date.

BJ Stacey

[Ken-ichi]

Regarding certainty, our current policy is that people should ID to
the taxonomic level at which they have absolute confidence. I don't
always do that myself so I'm being a bit hypocritical, but that's the
ideal. That said, as I pointed out the ID Please flag is still
working, so you could use that like you were using it before.

Regarding the reviewed checkbox, it should just save automatically.

[James Bailey]

Genus level observations don't seem to be getting research grade anymore: http://www.inaturalist.org/observations/1804558

I should have saved links to the other ones I found...

[Ken-ichi]

Genus-level observations will remain in Needs ID until they get to
species or are voted out of Needs ID. If voted out and they're below
family level, then they'll become Research Grade. Check out
http://inaturalist.tumblr.com/post/126691814973/changes-to-quality-grade

[James Bailey]

Ok cool, that was something I wasn't entirely sure about but good to have that clarified.

[Sam Kieschnick]

I like the update, but it'll take me a while to get used to, especially as I tell new users about the site.

One slight issue that I'm facing -- the power users among us are likely to clog up the "Needs ID" page...  Many of us post obscure plants or moths or insects that "need ID" and a new person will post a cardinal that also "needs ID," and it will get drowned in the umpteen hundred observations that the power users post...  I liked browsing through the "ID please" section to find some of the new users -- to welcome them, I liked to give them some guidance on their observations...

It sure would be nice to see a "observations from new users," but I'm a bit more biased towards making it more used by new folks (this has been discussed on another forum topic, so we don't have to get into it on this one...).

Again, I'm sure I'll get used to it, but I'll have to play with it some more.  :)

[lelliott]

I agree with Sam. I like helping people get ID's for things they don't know. For my own observations, I like having research grade observations, but it's a bonus. For people delving into new taxa or whatever, they probably NEED help narrowing an ID. So, Sam, try adding &has[]=id_please to the end of a query. I think that narrows the scope to observations that have been flagged id please. Not actually sure this is what the query does, I got it from one of Ken-ichi comments above. This doesn't prioritize new users, but maybe it will return the same types of observations as previously.

Lee Elliott
Columbia, MO

[Carrie Seltzer]

I find this thread interesting for learning how other people use the site. I think that's a great idea to further filter by Id please as Lee suggested. Needs ID and ID please are different fields in iNat (which you can see if you add .json to the end of an observation url--though you might need a json viewer extension on your browser to easily read it). I personally will spend a lot more time in Needs ID than I did in ID please but I'll do it for specific places or taxa I know. When I want to find new users, I search for observations that have no iconic taxon since that usually means they didn't know how/how important it is to attach their observation to iNat taxonomy. 

Carrie

—
Sent from Mailbox

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/rxohZR7QQ_k/unsubscribe.
To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Donald Hobern]

I just wanted to state how much I appreciate the new Needs ID model - it's suiting my needs perfectly and a vast improvement to what was already the best citizen science species observation site on the web.

Donald

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

[James Bailey]

One thing I've noticed is people ticking yes for Needs ID on an ID that is extremely unmistakeable. I can't be sure but I assume people think that it is correct to tick yes on this just because no one else has identified it even if there is no doubt on the ID itself. It means that even when people do agree with the ID it isn't becoming research grade unless people vote yes.

This is not an "issue" or "bug report', just an interesting observation on what is going on since implementation.

[Charlie Hohn]

Very true but I don't think it will matter much with the current system. The old Id please tag isn't as important anymore and it doesn't actually stop things from getting research grade. 

On Sunday, August 16, 2015, 'James Bailey' via iNaturalist <inatu...@googlegroups.com> wrote:

One thing I've noticed is people ticking yes for Needs ID on an ID that is extremely unmistakeable. I can't be sure but I assume people think that it is correct to tick yes on this just because no one else has identified it even if there is no doubt on the ID itself. It means that even when people do agree with the ID it isn't becoming research grade unless people vote yes.

This is not an "issue" or "bug report', just an interesting observation on what is going on since implementation.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[liselle...@gmail.com]

Hi!

Thanks for the changes and improvements made.

If you don't mind, I'd like to suggest that members be encouraged to participate in the identification of species by:

1. Sending email notification of observations to be identified.
2. Featuring an observation that needs to be identified. I suggest placing it on the right side of the home page(sharing a space with the map and place it above the fold). Pick an unidentified species that is rare and feature it for a week or until it is identified. Look into the older ones first before considering the newer ones.
3. Featuring a member and encouraging the community to help identify the observations made by that member. Pick an active member who has contributed interesting, rare observations from the wild. Pick also those who have actively participated in the identification of species.

Thanks and more power!

Liselle

On Friday, August 7, 2015 at 3:34:04 AM UTC+8, Ken-ichi Ueda wrote:

Hi folks,

In short: we're changing the way quality grade works for observations.
Instead of just "research" and "casual," we're

1) narrowing the scope of "research" to only include stuff ID'd at the
family level or lower,
2) introducing a "needs ID" state for observations that could be come
research grade but need more IDs, and
3) introducing an "unverifiable" state for observations that cannot
become research grade

We're also adding the ability to mark observations as "reviewed." This
will happen automatically when you add an ID, but you can also do it
explicitly for observations that you can't ID but don't want to see
again.

You can test these change at

http://gorilla.inaturalist.org
username: preview
password: 313phant

You can test out filtering by reviewed like this:

http://gorilla.inaturalist.org/observations?reviewed=false&quality_grade=needs_id

We'd love your feedback. The gorilla test site should allow you to
login and add IDs and stuff like you would on the live site, and it
will have no effect on the live site.


THE LONGER STORY

This all came out of some conversations here in the Google Group:

https://groups.google.com/forum/#!topic/inaturalist/F3HLBHQG0A0
https://groups.google.com/forum/#!topic/inaturalist/q4KiiVCwN-o

which we attempted to synthesize (after much haggling) into some new features:

https://github.com/inaturalist/inaturalist/issues/576
https://github.com/inaturalist/inaturalist/issues/611

There are a couple different needs that are being addressed here, most
of which are dealt with in detail in the above links, but the
overarching goal is to help people become better identifiers by
allowing them to focus on observations that need their attention.
We're also hoping these clearer delineations and filters will help us
build better tools focused on the experience of adding
identifications. We (the iNat team) are focusing on two main groups of
users at the moment: identifiers and new users. New users are
obviously important if we're going to achieve our primary goal of
connecting people to nature through technology, and our secondary goal
of creating enough data to be scientifically useful. We want to make
new naturalists, and we want them to make new observations. But in
order for them to get something out of iNat in the form of comments
and IDs, we need to expand and empower our core community of
identifiers. So this is a step in that direction, for the obvious
reasons of providing better filtering and the ability to ignore
things, but also because it helps clarify what we're all trying to
achieve here: helping people out by sharing our knowledge, and making
good data.

Anyway, looking forward to your feedback! We're hoping to get this on
the live site in the next week or two.

-ken-ichi

[M. Nicolai]

I am loving the new classification options. Just wanted to make sure I am using them correctly and ask a couple of questions about the data quality section. I have been looking for my observations where someone agreed with my id and as far as I know it cannot get any finer of an id, and then checking "no" on "still needs id". For observations where mine is the only id so far, but I am pretty sure they can't get any finer... I am leaving them as "still needs id". My thinking is this will allow them to be become research grade at some coarser level and then I can mark them as not needing id at that point. Is this correct?

I have been mainly using the observation photo uploader and then the batch edit section and there is question there asking if it is "captive/cultivated". When I check "no", is that information used somewhere? On the actual observation page it is reworded to say is it "wild/naturalized" which I would then expect to show "yes", but it doesn't. It is basically the same thing but asked the opposite way....if that makes sense.

Also love the suggestions mentioned earlier about featuring some rare observations. It might be neat to see ones featured that are new species or firsts for the whole site, but still need to become research grade to get added to species checklists.

[Charlie Hohn]

Another thing I think might be helpful would be an options 'random order' to the observations listed after an ID Needed query, or defaulting to 'older first'. I think it's important to look at the new things, where users can be given current feedback, but also to look at the old things, so they don't remain buried forever. A random option might offer a good mxi of that, or even maybe someday an algorithm that somehow chooses things of importance...

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at http://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

[Ken-ichi]

> checking "no" on "still needs id". For observations where mine is the only
> id so far, but I am pretty sure they can't get any finer... I am leaving
> them as "still needs id". My thinking is this will allow them to be become
> research grade at some coarser level and then I can mark them as not needing
> id at that point. Is this correct?

If you think *someone* could possibly identify your observation at a
finer level than the community ID, or there is no community ID, then
yes, I think you should leave the observation in "Needs ID," b/c as
far as iNat is concerned, the observation could benefit from more
identifications.

> I have been mainly using the observation photo uploader and then the batch
> edit section and there is question there asking if it is
> "captive/cultivated". When I check "no", is that information used somewhere?
> On the actual observation page it is reworded to say is it
> "wild/naturalized" which I would then expect to show "yes", but it doesn't.

In theory, when you apply the "captive / cultivated" flag from the
batch edit page and save those observations, it should vote "No" on
the "wild / naturalized" quality metric. I just tried that out and it
seems to be working. If that's not how it's working for you, can you
email a screenshot of your batch edit screen to he...@inaturalist.org,
as well as the URL of one of the observations you were trying to edit?

-ken-ichi

[M. Nicolai]

I just e-mailed the screenshots requested. I think it has always worked this way for me, but everything I have submitted so far was wild/naturalized so on the batch edit page with the radio form field captive/cultivated: yes/no/unknown options.... I have always set it to "no" just so it was not left at "unknown". Then I noticed the data quality info on the observation page said wild/naturalized: yes/no/unknown and they all still said unknown.

If I edit an individual observation (not the batch edit form) instead of a radio selection with the three options, there is just a checkbox to check if it is captive or cultivated. If unchecked I assume the status stays at unknown, but if checked it is set to captive/cultivated = yes (same as wild/naturalized = no). So using that form with just the checkbox (only two choices) there is no way to  indicate if something is wild/naturalized... if I am understanding it correctly??

[James Bailey]

A silly thing that still throws me off is how the Needs ID is opposite to the other data questions. For instance I'm used to seeing the green "yes" on all questions and it trips me up to see the red "No" when this is actually a positive thing.

Maybe Still needs ID? can be changed to something like Suitable ID? and the yes/no reversed so they are consistent with the other questions.

Just a thought.

[Charlie Hohn]

I found one case in which something isn't registering as research grade properly.  http://www.inaturalist.org/observations/110565 . It appears to be related to the taxon swap that happened. But it looks like it should still be research grade.