community identifications

[Ken-ichi]

Hey folks,

Apologies in advance for the novel. We're tinkering with how
observations are linked to taxa and could use your feedback.
Currently, *you* are in full control of the taxon associated with your
observation. Everyone else can add identifications, but your
observation isn't associated with a taxon until you agree with those
identifications.

However, this is a problem for people who don't realize they are
responsible for their own data, or for people who just add some data
and then abandon the site. The community puts work into adding
identifications, but the observations remain unlinked to taxa and
ineligible for research grade status. It's a bummer for identifiers,
some newbies, and especially for project owners who often recruit a
bunch of one-time users for an event who never return to confirm IDs.

What we'd like to do is introduce the idea of a "community taxon,"
which is sort of like the identification that the community of
identifiers agrees on. The algorithm we've come up with is a bit
convoluted, but roughly, it chooses the taxon that more than 2/3 of
identifiers agree with. You can see an example of this by appending
test=cid to any observation added in the last week, e.g.

http://www.inaturalist.org/observations/505772?test=cid

(If you want the gory details on the algorithm click the "About" link
beneath the community ID, or wade through the even more crazy diagrams
at https://github.com/inaturalist/inaturalist/issues/88). So far we're
just calculating the community taxon without affecting any existing
data or quality grade, but I think what we want to do is set the
community taxon as *the* primary taxon affiliation for an observation,
*unless* you opt out. This would mean that if you like the current
system of personal control (as I do), you would have to set a
preference saying you don't want to automatically accept the community
taxon as *the* taxon for your observation.

Assuming my verbose explanation was comprehensible, what do you guys think?

-ken-ichi

[Charlie Hohn]

I'm in support of the idea. the photo and the original user's ID aren't changed so I don't see it too much as a control issue... if I was certain the community had misidentified something or had another concern about one of mine I'd just delete the observation.

[Kent McFarland]

I like it Ken-ichi. I think this will help clean up a lot of hanging chads. I wonder if perhaps also we could consider adding some editing abilities for project curators? Or for vetted curators. By that I mean a group of curators that have been given higher powers. I say this as I have come across obs that are clearly misidentified or "over-identified" (meaning that it might be that species but impossible to tell from photo given other closely aligned taxon), but all agree on an ID so it is research grade. I can comment on those and it is often followed up on by someone, but it would be nice if some curators had higher powers to clean things up a bit. Just a thought.

Kent

____________________________

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

   

--
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/groups/opt_out.

[Charlie Hohn]

I second the idea that it would be neat to have a reputation or certification system 'someday' that would allow for experts in given fields to have more weight to their observations and IDs. I don't mean to restrict it to only 'professionals' as some so called 'amateurs' are experts and should qualify too... but it would have to be a pretty selective group.  For instance, I think there could be value in uploading some plant survey from our Heritage department and our botanists, if we decide we want to do that, and those seem like they could definitely qualify as 'research grade' but without photos won't ever get there.  Basically we have tons of data on non rare plants that just never gets used and just sits in notebooks. On the other hand, there are some plants I know I am uncertain of, for instance, and if I put in a bird without a photo, and it's not a chickadee, the observation should be as far from research grade as possible :)  So maybe to make something without a photo research grade would require a checkbox, or maybe 'research grade' is the wrong term to use for these.  But anyway, this is a bit off topic so I'll stop here.

[Mikael Behrens]

Ken-ichi and all,

I think this is great! It puts more faith in the iNaturalist community which I think deserves it.

In addition to the situations you listed, a community ID will help observations that languish in an inactive taxon identification or a Something identification with unaware owners. The community ID makes it more worthwhile for experts to surf non-research-grade observations and add identifications. And it makes adding multiple agreements to an identification have more of an effect.

About the reputation or certification idea, maybe a user's certification could be incorporated in the Community ID algorithm somehow. For example, a user could be certified as an Expert in a certain taxon and maybe also in a certain geographical area for that taxon. (For example, user greglasley would easily be an Expert in Odonata in Texas.) Then that user's identifications would have more weight in the Community ID algorithm if the observation was under that taxon and in that location. What do you think?

Mikael Behrens (mikaelb)

Austin, TX

[Chris Brown]

Ken-ichi,

 

Sounds like a great idea! I especially like having the opt out option, allowing the user to maintain control over their observations should they choose. I'm less enthused about the "expert weighting" system though. It seems to detract from what iNaturalist started out to be...but understand the want/need for it. Even experts can make incorrect ID's from photographs, especially in the invert world.

 

Chris

[Charlie Hohn]

Invertboy, that's a very good point. I guess my best response is that this is true for experts recording data for agencies and consulting firms too.  Here, there is the potential for others to correct a wrong ID by an 'expert', and if that person truly knew the natural world and was humble and sensible they would back off to genus or adopt the community ID.  I don't think 'expert' should ever outweigh the community.  I think other than the research grade system, there are some issues with the current system including students who are learning putting in IDs that are agreed to by other students who also are just learning, but with enough community involvement that would be overridden. Maybe a better way to solve the problem would just be to change 'research grade' to 'community consensus' or something, and have another grade for people with a high enough reputation or whatever, that could be shared with other agencies or not, depending on what the case may be.  It isn't an essential feature, if I want all of Kent's tree IDs including those without photos (if he had any) I could still download those.

Another related thought... I know tagging 'something' observations with a very basic type like 'dicots' or even to genus can cause problems with the observation reaching research grade if I forget to go back and remove my ID if someone else has a more precise one. I wonder if it would be possible to set it up so that those types of IDs don't count as votes against the identification.  If I see something and know it is a pine but not the species, but two other people agree it is Ponderosa Pine, it doesn't seem like me saying it is 'pine' should decrease the weight of that ID.  On the other hand, if they identify it as pine and I think it is a cedar and mark it as a cedar, I would think that should count as a disagree.  Perhaps someday there could be a field for 'disagree with given ID but don't know real one' and a field for 'this is a speculative suggestion to give you something to work with, don't let it count towards or against research grade'.  But then things get much more complex.

Sorry about the rambling post.

[Charlie Hohn]

one other note, maybe there should be a way to mark an observation so that I see that I looked at it but didn't know what it was, so I wouldn't see it again.  But maybe that creates a database nightmare matrix.

[Charlie Hohn]

Ahh! Sorry for the triple post but I see that the new system actually accounts for my concern about 'higher' ranks being put in first!

[AfriBats]

I remember that we already had this discussion, and I agree this is a very important issue that can hopefully be addressed soon:

If there's a temporal refinement of the identification from, let's say, the first person giving an ID at the level of order, the next person an ID at family level etc. until the observation is eventually IDed at species level, this shouldn't prevent the observation from becoming RG. I see loads of observations where initial IDs at a higher taxonomic level are probably simply forgotten (or where those providing these initial IDs would only feel comfortable at the given level), so introducing some mechanism ignoring previous IDs, if in the correct taxonomic branch, is very much needed!

[AfriBats]

I think this issue should be addressed independently of the community ID.

[Belinda Lo]

I think it's a great idea too! Will we be able to opt out for particular taxa or individual observations?

I agree with Chris that even people who are considered experts can make incorrect ID's. And in my mind, the true experts are the taxonomists who create the literature on which the rest of us base our ID's on.

Belinda

 

[Ken-ichi]

Thanks for the input, everyone. Good to know most people seem on board.

Belinda, you will be able to opt-out for individual observation or for
all your observations, but I'm not planning on making a taxon-specific
opt-out.

Regarding ID refinement, yes, the algorithm (which muir and loarie
came up with, btw) accounts for that. If the sequence of IDs looks
like

Mammalia
Mammalia
Homo
Homo sapiens
Homo sapiens

then the community ID will be Homo sapiens. Disagreement can still
occur by adding higher level IDs *after* more specific IDs, though in
trying to come up with an example I found a bug! Will fix.

Reputation / certification: not going to deal with that in this work,
but maybe some day. If we do implement such a system, reputation will
be earned, not bestowed from on high.

-ken-ichi

[Charlie Hohn]

Sounds great!

[AfriBats]

What I'm suggesting is that this temporal-hierarchical succession of IDs, if correct, should not prevent an observation from becoming Research Grade. For instance, in this observation
www.inaturalist.org/observations/505151
four people would either need to remove their ID "Acacia", albeit probably correct, or agree on "Acacia retinodes" before the observation turns into Research Grade, right?

The important point here being whether or not the original observer eventually added the species level ID. If so, previous IDs of _correct_ higher IDs should not matter.

[Chris Brown]

Awesome, I like the "earned" vs "from upon high" approach.

[AfriBats]

Gee, difficult to explain, let me try again with the example:

I suggest that this should turn into RG
Mammalia [original observer]
Mammalia [user 1]
Primates [user 2]
Homo [user 3]
Homo sapiens [user 4]
Homo sapiens [original observer]

while this should not turn into RG
Mammalia [original observer]
Mammalia [user 1]
Homo [user 3]
Homo sapiens [user 4]
Homo sapiens [original observer]
Homo [user 5]

assuming that User 5 objects to the ID as "Homo sapiens" in the latter case.

Note that all of this concerns whether and how an observation turns into RG independent of the community ID.

[Charlie Hohn]

AfriBats, I tested that observation using the new community ID protocol paste in(?test=cid ) and the species came up as the ID! So it sounds like your issue, which is by the way a very valid one, would be solved.

[Mikael Behrens]

One more (maybe minor) thought. In the user interface, should the Community ID have an "Agree?" link? Currently it does, and this might encourage people to agree with a more general ID without considering more specific ones. For example, see this observation:

https://www.inaturalist.org/observations/505063?test=cid

Also, notice that the Royal Tern ID (which I believe is correct) does not have an "Agree?" link.

Mikael Behrens

[Ken-ichi]

I feel like the tendency to click the "Agree?" links instead of taking
the time to make sure you're being as specific as you can is already a
problem, but a minor one, so I'd rather wait and see if it becomes a
more serious problem with community IDs. That page doesn't show the
"Agree?" link next to Royal Tern for you because that's already your
ID. I can see it because I haven't added an ID.

[Mikael Behrens]

Thanks, that makes good sense to me!

Mikael

[Scott Loarie]

Jakob (Afribats),

Your first example:

Mammalia [original observer]
Mammalia [user 1]
Primates [user 2]
Homo [user 3]
Homo sapiens [user 4]
Homo sapiens [original observer]

would get a community ID of Homo sapiens (rank species), because as in
the obs you mentioned
http://www.inaturalist.org/observations/505151?test=cid
these coarser IDs (ancestors) were added before the finer IDs
(descendants) they don't count as disagreements with the finer IDs -
i.e. they don't add weight to the denominator of the (votes
for)/(total votes) ratio so (Homo sapien + Homo sapien) / (Homo sapien
+ Homo sapien) = 2/2 > 2/3 means Homo sapien would be the finest rank
with >2/3 votes and would take the community ID.

However, if these coarser IDs are added after the finer IDs, as in
user5 in your second example, Homo would be implied to mean a
disagreement with the finer ID of Homo sapien so there would be 2
votes of Homo sapien out of 3 total votes (Homo sapien + Homo
sapien)/(Homo sapien + Homo sapien + Homo) = 2/3 !> 2/3 so the
community ID would go to Homo (rank genus) because (Homo + Homo + Homo
sapien + Homo sapien)/(Homo + Homo + Homo sapien + Homo sapien) = 4/4
> 2/3 as the lowest rank with >2/3 votes

while this should not turn into RG
Mammalia [original observer]
Mammalia [user 1]
Homo [user 3]
Homo sapiens [user 4]
Homo sapiens [original observer]
Homo [user 5]

But the issue of research vs. casual grade is sort of a different
issue. With the community ID in place, observations with
photos/dates/locations and no captive etc. flags will be Research
grade observations of the taxon identified by the community ID (as
opposed to the taxon identified by the user's ID as the system is
now). But in both systems observations can be research grade
regardless of the rank of the taxon (e.g. an observation could be
research grade at rank kingdom = Animalia). If I understand, your
concern is that observations with ID coarser than species maybe
shouldn't be research grade? I'm not sure I agree with this because
(a) there are uses to data not ID'd to species, for example a study on
newts, might want all Taricha observations regardless of whether they
are Taricha torosa or Taricha granulosa, (b) GBIF includes records not
ID'd to species which make me think these data are of interest, and
(c) by usng the rank filters, people can filter out the rank at which
obs are ID'd independent of Research Grade, so you could for example
only look at RG bats ID'd to species or finer:
http://www.inaturalist.org/observations?quality_grade=research&taxon_id=40268&hrank=species

But I do agree with you that we should come up with a better way of
highlighting observations with coarse IDs that can be ID'd to finer
ranks and getting specialists to work on improving these IDs. This can
be done with existing filters, for example as a bat specialist you
could filter all observations with community IDs in the order
Chiroptera having photos that are of rank genus or coarser and try to
improve these IDs. But maybe there's a way to highlight this use case
to specialists or as Charlie has mentioned, come up with a way for
people to mark observations as reviewed (e.g. I've looked a this genus
level bat observation several times, but I know it can't be resolved
to species from the photo so would prefer it doesn't get returned in
my search).

Another feature I think will be useful to build would be user mentions
on observations https://github.com/inaturalist/inaturalist/issues/179
- as is, you might not alone be able to move the community ID of a bat
observation from genus level to species (would need >2/3 votes), but
if you could alert two other specialists to also add IDs, it would get
the 3 species level IDs needed to push the community ID down to
species level. Mentions (e.g. 'I bet @charlie could ID his to
species') might help (in the meantime just sending people messages
would do the trick too).

> --
> 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/groups/opt_out.

--
--------------------------------------------------
Scott R. Loarie
www.stanford.edu/~loarie
Email: loa...@stanford.edu
Phone: 415-278-1220

Post-Doc, Dept. Global Ecology
Carnegie Institution for Science
260 Panama Street
Stanford, CA 94305

Lecturer, Dept. Geography
561 McCone Hall
University of California
Berkeley, CA 94720

Co-director, iNaturalist.org
California Academy of Sciences
55 Music Concourse Dr
San Francisco, CA 94118
--------------------------------------------------

[Charlie Hohn]

The alert feature would be awesome and I'd definitely use it, maybe to the point of harassing Kent, Chris, and BotanyGirl too much :)

On Tuesday, January 21, 2014 3:03:20 AM UTC-5, Ken-ichi Ueda wrote:

[Ry Beaver]

I like this concept - I have been working through countries (Well I have gone through Aust a couple of times and started the UK) trying to convert as many obs into RG as I can - and it has been a bugbear where I know it is a specific species and can tell from the photo but the originator has gone AWOL or hasn't agreed with my suggestion as they are not actively checking iNat or don't realise they need to do anything.

[Rob Curtis]

Wow, this is a great cutting edge discussion. I believe you are resolving longstanding issues I have had with a) knowing that the ID is wrong, but not knowing what actually is, and b) believing the observer took the ID too far.

However, what of the cases where the observer knows the actual species, but records the species during a time when it cannot be positively identified through photos? For example, the basal rosette of a known flowering plant or insect eggs of a known source. Should there be a way to elevate these photos to RG as well?

And finally, what about cases where there is a record without photo (or sound clip) of a species that you may have also observed (and know exists) at that location, but may be nearly impossible to document. Can agreeing with these records eventually take them to research grade?

[AfriBats]

Scott:

I agree that the RG issue is to some extent conceptually different from the community ID.

What I currently don't understand is which of the systems will determine whether an observation becomes RG or not: will this be based solely on the community ID, the traditional ID system, or both? With both I mean that the 2 systems could independently lead to an observation becoming RG.

With the 2nd example I simply wanted to illustrate the point that the ID of "Homo" by user 5 means something different from "Homo" by user 3 since it is added AFTER 2 users IDed the observation down to species level.

The question whether or not IDs at taxonomic levels higher than species should be allowed to turn into RG is still another topic. Don't know whether we should include it here. I would be indeed in favour of restricting RG to species level.

Finally a more general comment: I would try to keep the system as simple as possible. Introducing additional check boxes, alerts etc. that require or ask for the input of peers might be stretching the community involvement too much. I guess that many of the discussions we're having in this forum are driven by quite a few power users, and a lot of iNat users probably do not even know that the forum exists.

Sorry for these varied thoughts, some of them being rather unrelated to the main discussion here!

Cheers, Jakob

[Scott Loarie]

HI Rob - it seems like a reputation system would need to be in place
to build a system for elevating observations without shared evidence
to research grade.

While research-grade observations based on reputation is probably a
bit down the road, it seems like a first iteration of a system for
earning reputation might not be too far off - e.g. first person to
propose and ID that eventually gets backing by the community earns
reputation.

[Scott Loarie]

Hi Jakob,

I agree with you about keeping the interface simple.

Re: what determines Research Grade, in both systems the existing data
quality assessment metrics will stand:
Community-supported ID?
Date?
Georeferenced?
Photos or sounds?
Is the organism wild/naturalized?
Does the location seem accurate?
Does the date seem accurate?

But the meaning of 'Community-supported ID?' will depend on whether
the user opts-out of community IDs for the observation or not

If the user does not opt-out:
'Community-supported ID?' will depend on whether a Community ID exists
(an ID with >2/3 support and at least 2 votes) and the taxon
associated with the observation will be determined by the Community ID

If the user opts-out:
'Community-supported ID?' will depend on the current rules (more
agreements with the User's ID than disagreements) and the taxon
associated with the observation will be determined by the User's ID

Re: encouraging finer IDs, I think you make an excellent point about
the need for a system to alert specialists to review observations that
need finer IDs. But since you're right that with Community IDs will
make nearly everything with photos and at least one other ID Research
Grade at some rank, people will need to use other filters rather than
'quality grade' to find observations they can IDs contribute to.

For example:
http://www.inaturalist.org/observations/507464?test=cid
jimsteamer's ID: Butterflies and Moths (Order Lepidoptera)
jakob's ID: Eggflies (Genus Hypolimnas)

Under Community IDs, this observation would b Research grade at rank
order. Because of the 'at least 2 votes' rule, this observation needs
another ID of Genus Hypolimnas to move the Community ID from Order
Lepidoptera to Genus Hypolimnas. How can we ensure that this
observation is brought to the attention of someone who can provide
this ID. I see three options:
1) get specialists comfortable with learning the Observation search
filters so that they can search for observations in their specialty
taxon/place with 'highest rank' set to something coarse and 'w/
photos' checked. This will return observations they can hopefully help
with.
pros: tools already exists
cons: not readily apparent how to do this and might be hard for
specialists not familiar with the site to get used to these filters

2) build something similar to the 'ID please' page, maybe 'needs ID'
that does this automatically, returns all observations with photos or
sound ID'd with rank coarser than species with easy controls for
setting the place/taxon
pros: might be an easier tool to get specialists to learn with fewer
buttons than the raw observation search filters.
cons: redundant functionality for something that can already be done
with search filters, might be confused with ID please, photos/sounds
might be insufficient to provide finer IDs

3) make it easier for people to alert specialists of observations they
can ID. For example, if you know that I have good African butterfly ID
skills, you could send me a message asking me to take a look at obs
507464 - mentions '@loarie would know this' would keep these messages
in the observation comment/ID stream.
pros: specialists just have to check their dashboard - don't need to
navigate any filters
cons: doesn't require motivated specialists, unmotivated specialists
might be annoyed and feel flooded with ID requests

[Christopher Tracey]

Regarding specialists, would there be a way for a specialist to flag an observation as only identifiable to a specific level (eg. Genus) due to lack of appropriate features or data? This would prevent observations from reappearing in specialist's lists of species to work on.

- Chris

[Charlie Hohn]

I think individual flags make sense but I'm not sure about an overall flag.  For instance I used to do vegetation mapping in the Santa Monica Mountains using binoculars and now if I have a location and a good picture, I can identify things (chaparral shrubs especially) that most people couldn't.  On the other hand, a slightly blurry picture of a bird and I'd be lost. I might flag a bird, and then miss a plant I might potentially know.  So I'm not sure what makes sense.

[Kent McFarland]

There are plenty of things, especially arthropods, in which it is impossible without a specimen to ID to species level. I see things "over-IDed" regularly. 

Kent

____________________________

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

   

[Charlie Hohn]

Do you mean by the original poster or by subsequent observers? The former I usually give the benefit of the doubt that they might have seen more, but for the latter, yeah, it definitely happens, especially by less experienced students... some of the school groups in California were doing it, for instance. (sometimes to the wrong ID). But it seems rare

[rcu...@summitmetroparks.org]

I can see it both ways. I definitely agree with Charlie. I have stated
already (and maintain) that there are times when RG should be accessible
with ONLY a date and location (ie. without a photo or recording), if a
majority can agree on its presence at that location and time. However, I
have also seen species which I strongly BELIEVE are either wrong or
identified too specifically. Yet I also know from experience on iNat that
just because I may only be able to ID a plant by flower
characteristics. . . there may be someone else who can do it by bud or
pubescence or familiarity with the options in the region. I am afraid
there are some who want to maintain a monopoly on ID rights (which project
curators already have). I believe that the proposed Community ID scoring
system, though perhaps not perfect, will allow a balances means for both
concerns to be addressed.

From: Charlie Hohn <naturalis...@gmail.com>
To: inatu...@googlegroups.com
Cc: kmcfa...@vtecostudies.org
Date: 01/23/2014 09:29 AM
Subject: Re: [inaturalist] Re: community identifications
Sent by: inatu...@googlegroups.com



Do you mean by the original poster or by subsequent observers? The former I
usually give the benefit of the doubt that they might have seen more, but
for the latter, yeah, it definitely happens, especially by less experienced
students... some of the school groups in California were doing it, for
instance. (sometimes to the wrong ID). But it seems rare

On Thursday, January 23, 2014 8:12:26 AM UTC-5, Kent McFarland wrote:
There are plenty of things, especially arthropods, in which it is
impossible without a specimen to ID to species level. I see things
"over-IDed" regularly.
Kent

____________________________


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

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/Fh7Gxfg1EpQ/unsubscribe.
To unsubscribe from this group and all its topics, 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/groups/opt_out.

This message and any response to it may constitute a public record and thus may be publicly available to anyone who requests it.

[Ken-ichi]

Chris, you can "flag" an observation as only being identifiable at the
genus level by identifying it at the genus level. If enough other
people agree with you, then that will become the community ID.

Regarding observations without media becoming research grade, that's
not going to happen without a reputation system, so let's talk about
it if/when that happens.

[Charlie Hohn]

One other question... how will this work out with the maps? I look at them a lot since i'm interested in the range of different plant species. If I see a burr oak and mark it as such but have a bad photo so people can't tell and just mark it as 'oaks'... I agree it shouldn't be research grade but I kinda think it should still show up on the burr oak map within iNat. I could of course withdraw from research grade but I might want to be able to see things others think are burr oak too. Anyone else have thoughts on this?

[Kent McFarland]

I think you'd want them to be a different icon or something if that is the case. It would be neat to see map with research grade as one icon and others as another. 

Could push things further to research grade as it might get more people to help verify. 

Kent

____________________________

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

   

[Charlie Hohn]

Yeah, that would be awesome. The more icons the better (at least from a user standpoint, not sure of the programming side).  For instance it would be neat if observations I'd made, or observations I'd helped ID, were a different color or symbol too.  But hardly necessary.

[Kent McFarland]

Yeah, indeed Charlie, anything I ever say is NOT from a programmer view for sure! I bet Ken-ichi has groaned over and over at some of my comments on here! Sorry Ken-ichi! I always like when he says to me something like -  that would be great, you should write some code for that. :)  

Kent

____________________________

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

   

[Charlie Hohn]

Yeah, every once in a while I think I should learn some programming, but I never get past the silly names (Ruby on Rails? What does that even mean? Did I get that right?)  Well, maybe it will be a project some day but I feel like the programming stuff would update faster than I could teach it to myself. Also it would might not be a good idea for me to start coding changes to iNaturalist because I might get carried away :)

[AfriBats]

Excellent idea, I'd be very much in support of this! Would suggest only 2 different icons (RG and non-RG): I'm frequently going by the (bio-)geographic pattern, and using different icons on the map would help to jump to those outliers no one has verified (or where there's disagreement and thus preventing the obs from becoming RG).

[Ken-ichi]

Re mapping, observations will be mapped based on the community taxon
if it has been set the observer hasn't opted out. We definitely need
better cartography and better maps in general to add context about
quality grade, wild/captive, etc, but that's a separate issue (one
that we're working on with some partners in Spain).

[Rob Curtis]

There are two occasions when one might not make an ID to species/subspecies. The first (and the one you address specifically with your equation) is when the more specific ID is suspected to be wrong. However, I (and I suspect others) often only feel comfortable going to genus or species (rather than species or subspecies, respectively) because of our own limitations. In this case, it should not necessarily take away from a more specific ID just because some or a majority of others cannot go as far as others.

[Charlie Hohn]

Yeah, i feel like there is an etiquette now not to id something to a lower taxa level just because you don't know either way... for instance I  skip subspecies IDs if  I don't know them. It probably needs to be spelled out that way or something though.

[Scott Loarie]

Hi Rob and Charlie,

You raise a good point that if an observation has a community ID at
some rank (say species) and you are only able to ID it to a coarser
rank (say family), adding such an ID at this coarser rank would be
implied as a disagreement. So I guess the etiquette should be that if
you can only a contribute a coarser rank and don't explicitly want to
disagree with the IDs at finer ranks you should probably not add an
ID. This should probably be spelled out in the Getting Started Guide
http://www.inaturalist.org/pages/getting+started

[Ken-ichi]

Alrighty folks, just about ready to roll this out. I wrote a blog post
describing the new system and I'm running a banner on the home page
for a week before releasing this:

http://inaturalist.tumblr.com/post/75092437669/community-identifications

I want to make sure control freaks like me have time to opt-out if
they want to. As always, if anyone has questions or comments, please
fire away.

[Charlie Hohn]

Awesome, I look forward to it!

i often go straight to my bookmarked queries rather than the front page so if it is easy and you were really a control freak you could send a notification too... but definitely isn't necessary

[Kent McFarland]

Ken-ichi,

I will put a note on the Vermont site for it too. Thanks for your hard work on this.

Kent

____________________________

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

   

[Paul Hamilton]

Awesome, this will help a lot. The only pitfall I see is that it would sometimes give too much weight to casual identifiers. For instance, a lot of frogs look like tree frogs, but most that look like tree frogs are not. So, anything that looks like a tree frog gets identified as such, and then the "agrees" come easy.

One possible way to address this would be to add a "disagree" option alongside the "agree". If you know that the "easy" identification is not right, but you still don't know what it is, you could click that. And then maybe that identification doesn't get weighed in the community-supported ID algorithm?

In any case, a great step forward.

Paul

[Scott Loarie]

Paul,

In that example you'd just need to add an ID of frog after the treefrog IDs to back the community ID up from family to order

Sent from my iPhone

--

[Lynn Watson]

Excellent stuff.

Is there a way to flag an observation belonging to someone else (the obs does not have a Help Id) that it needs some help from others, to get the number of votes up to changing the obviously incorrect id. 

[Charlie Hohn]

I think in general it would be nice to have a mechanism to 'churn' up old non research-grade-to-species observations. But I am biased... I made a ton of observations when iNat wasn't being used much outside of California, and while some of them have poor photos and will never get research grade, some definitely could if people ever found them...

[rcu...@summitmetroparks.org]

I agree Charlie.  I noticed that PN brings records back to the surface when any changes occur and I really like it.  So if I narrow my two year old insect record down to a weevil tomorrow, my change would pop it back to the top of the pile and potentially be sent to any weevil subscribers.  This would give iNat a much more dynamic front-end flow and let everyone see who is doing what behind the scenes.

On a similar note, I have seen several cases when photos were swapped out on records where I had made previous IDs and I never knew it happened. I feel that any change, including deletion of records (if possible) should be signaled to all invested parties.

Rob

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/Fh7Gxfg1EpQ/unsubscribe.
To unsubscribe from this group and all its topics, 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/groups/opt_out.

[Scott Loarie]

You can see a static display of recent comments if you go here:
http://www.inaturalist.org/home and click 'Comments by everyone', but
its slow to load, and I agree I also like the PN front page which has
the dynamic feed of comments mixed in with comments and favorites.

[Charlie Hohn]

I was wondering if the community ID would take effect first thing this morning!  But, nope! :)

[Ken-ichi]

Nope, but I am rolling them out today. First need to deploy the
changes, then start migrating all the existing observations, then deal
with the ensuing chaos.

On Fri, Feb 7, 2014 at 6:02 AM, Charlie Hohn

[Charlie Hohn]

Yeah, I figured it was going to cause some database chaos.  Will the site go down?

[AfriBats]

Hi everyone

Great to see the community ID coming into effect! I guess we'll encounter special cases where the system might need to be fine tuned. What about this one:
http://www.inaturalist.org/observations/498498

I think the Community ID should be Bubo africanus rather than Bubo. Any thoughts? Cases like this (outdated IDs due to taxonomic splits) might be a rather common issue, and previous IDs of the now inactive taxon shouldn't be counted as a disagreement.

Cheers, Jakob

[Ken-ichi]

In the case of swaps maybe they shouldn't count as disagreements, but
in a split like this (http://www.inaturalist.org/taxon_changes/4127)
where we can't really say that an ID of an inactive concept like

http://www.inaturalist.org/taxa/20092-Bubo-africanus

is equivalent to an ID of active concept like

http://www.inaturalist.org/taxa/204470-Bubo-africanus

I think they should could as disagreements. Or maybe they should count
as identifications at a rank higher, like genus, but that just seems
even more horrifically complicated than the algorithm already is.

[AfriBats]

Actually I thought of counting them as the next higher rank, i.e. genus in this case, but if this complicates things too much, then let's forget it. (or add an irritating, blinking red button next to inactive taxa to alert people that something should be taken care of there...)

[AfriBats]

Maybe I've missed something, but I thought the temporal succession of IDs would matter in the Community ID (older but taxonomically higher IDs not preventing newer & taxonomically lower IDs). How about this one:
www.inaturalist.org/observations/521437

The older ID is that of the genus, the newer down to species. I'm sure about the genus ID but don't know the species. I don't wont to prevent the observation from showing up as the ID down to species level, so would I need to remove my genus ID for that?

[Charlie Hohn]

Well, it still wouldn't be research grade to species until someone else agreed with it. Though I do see the point that it stops it from displaying as the species too.

[AfriBats]

Hi Charlie, note I'm not talking about whether or not it should be RG at species level but how the observation displays. Currently, it won't show up as Charaxes brutus, e.g. on the map of that species:
http://www.inaturalist.org/taxa/126296-Charaxes-brutus

[Charlie Hohn]

Yeah, I definitely see what you are saying. It might cause database problems if it was research grade to genus but displayed by species? But it does seem like if the user enters a species name after someoene else IDs to genus it would be best for it to display that species name.  Maybe it's hard from a database standpoint. In the short term, it's not a perfect solution, but the user can opt out of community ID for that one observation. Beyond that I guess we will have to see what the site admins say.

C

[Scott Loarie]

The tradeoff is whether, for a given amount of information (IDs),
observations are cataloged more conservatively at coarser ranks with a
reduced incidence of wrong IDs (e.g. higher percentage research grade
IDs) or less conservatively at finer ranks with a larger percentage of
potentially bad IDs (e.g. fewer research grade).

As charlie mentioned, the owner (snidge in Jakob's example) can always
shift to the less conservative option by opting out of CIDs for an obs
(that would make this a non-research grade observation of Charaxes
brutus).

In Jakob's particular example, without adding any species level
insight, Jakob also can shift to the less conservative option by
removing his ID of Charaxes. Its true that non-obs owners don't always
have the power to do this under the current CID system. For example,
there's nothing I could do to make this a non-research grade
observation Charaxes brutus. I haven't totally though this through but
my gut tells me that without any new information (IDs), the current
system is appropriately conservative - e.g. less wrong data even if
its coarsely ID'd. But I definitely see Jakob's point.

The best solution would be for Jakob's example to be cataloged at the
species level (Charaxes brutus) AND be research quality. And just one
person familiar with Charaxes brutus needs to chime to make this
happen. How can the community encourage this process of getting the
right observation reviewed by the right expert? This is something that
Bugguide seems to do quite well, and something I've always marveled
at. Will some new functionality make this happen? If so what? Or is it
a logistical issue, and if so what can we as a community do to come up
with a protocol to filter these observations to regional and taxonomic
experts somewhat systematically (AND get them to respond!). Or will
this just come naturally when iNat has a big enough community, so
should we just concentrate on that?

> --
> 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/groups/opt_out.

[AnitaG]

Yes, I actually came on over to post about this.  Can the Community ID system be made more taxon-swap-aware?  Right now, if you post a obs under 1 ID, someone agrees with that ID, and then the observation gets transferred to a different taxon because of a taxon swap, then the Community ID suddenly reads as if there is a disagreement as to the ID.  Here's my example: https://www.inaturalist.org/observations/537416: I posted this as Dolichomia olinalis, a fellow user agreed, and then a curator came along too and said, “Oh, that genus was changed; I’d better fix the database.”  So he implemented a taxon swap to Hypsopygia olinalis, & I agreed to it for my observations.  But now the community ID on my obs reads Pyralini.

Best,
Anita

[Matthew Muir]

Not sure if this is a bug, or working as intended: http://www.inaturalist.org/observations/63379

1 ID at SubFamily, 2 IDs at Species, yet Community ID is at SubFamily:

Shouldn't Community ID = Species?

[Scott Loarie]

Whats happening here is that the genus Parrhasius is not grafted to subfamily Theclinae but rather directly to family Lycaenidae (e.g. Thclinae does not include all of its descendants). This is something a curator can fix by editing genus Parrhasius and changing the parent from subfamily Theclinae to family Lycaenidae

These kind of issues happen alot when people add intermediate ranks (e.g. subfamily) not in dataproviders like COL but don't complete the curation of the descendants by moving them all to the proper parent. Would be awesome to have better tools for doing this.

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.