GPS Coordinates
Lirpa
The site looks quite promising. I can't wait to see what it will look
like once it have more observations!
Couple comments and concerns:
1. I have the GPS coordinates for some things that I saw and would
like to input that instead of searching for their location on the map.
2. I also think that it would be nice if I could access the add
observation option from any page on the site.
3. It would also be nice if I could click on the location flag and
then have the option of adding my own observation directly.
4. Can you link observations, if a group of people see the same
thing?
5. It will also be difficult to confirm a sighting of something if
there aren't any pictures. How are you going to screen for accuracy?
Good job!
Ken-ichi
Thanks for signing up! Here are some replies to your excellent questions:
> 1. I have the GPS coordinates for some things that I saw and would
> like to input that instead of searching for their location on the map.
When you make a new observation, you can click the "Add additional
data" link and you will see latitude and longitude fields. You can
enter GPS coordinates there, but they'll have to be in decimal degrees
(negative values for degrees longitude west). If you have GPS
coordinates in other formats like degrees/minutes/seconds or the
dreaded degrees/decimal minutes, you'll need to convert them using
something like this:
http://www.csgnetwork.com/gpscoordconv.html (hideous, but functional)
So W 121 deg 33.003' becomes -121.55005.
> 2. I also think that it would be nice if I could access the add
> observation option from any page on the site.
Good point. I've been meaning to add that...
> 4. Can you link observations, if a group of people see the same
> thing?
You can't right now, but it's a feature we've discussed. How do
others feel about this? Do you want to be able to say "I saw species
X at time Y in place Z with April, Ken-ichi, and Nate?" How would you
feel about the observations that others have added in this way showing
up as your own if they claimed you were there too? Maybe you could
limit that ability to your contacts, so only certain people could add
to your observations.
> 5. It will also be difficult to confirm a sighting of something if
> there aren't any pictures. How are you going to screen for accuracy?
Very true. Our philosophy is not to screen for accuracy, but to
enable people to screen for it themselves. We haven't quite gotten
there yet, but here are some possibilities:
- filtering observations by whether or not they have photos
- allowing users to suggest (and possibly confirm) IDs, so you could
say, "only show me observations that have been ID'd the same way by at
least 2 people"
- building this database of confirmations and suggestions through
games. Imagine a game where you're simply shown an observation
without the species (a photo, a place, a description) and asked to
make the ID. You could even have a competitive version similar to the
Google Image Labeler (http://images.google.com/imagelabeler/).
Without a photo it's definitely going to be harder to confirm.
Extensive descriptions certainly help, but there's also potential to
leverage the power of social networking. Let's say you have a private
list of "trusted contacts" or something, and you can sift through
observations based on degrees of separation, e.g. "only show my
observations made by people I trust and the people they trust." Just
a thought.
Data quality is a thorny issue. We don't want to create barriers to
creating observations, but at the same time we want to incentivize the
creation of accurate data. We'd love to hear more about what you all
think about it, and any other ideas you might have for addressing it.
-Ken-ichi
George
over for the site, I'm now leaning more towards your basic vision of
not screening. The photo screen-out would be an incredibly difficult
standard. I'm reminded of the recent wolverine sighting headlines.
Most wildlife people are quoted as saying none have been seen since
one was shot (!!) in 1922. I am 99% certain there have been
significant wolverine sightings in both Yosemite and Sequoia Kings as
recently as 5 years ago. Years ago a wildlife biologist friend of mine
saw a Mt. Lyell Salamander a fair distance south of its known range.
That sighting was not accepted by the UCB herp. folks because he
didn't put it in a jar. Of course, they're all over the place (sort
of) down there and a number of collecting trips have now established
that for the UCB folks.
Years ago I had a researcher dismiss my sightings of water shrews at
high altitude because, he said, they don't occur that high. Grrrrrrrr!
But there there's the flip side of reporting "a family of wolverine"
as was once reported to me. They were Marmots, alas.
So that would argue for maybe an optional (!!) self-reporting of
experience on some sort of scale of 1 to 5; a scale of how sure you
are of the observation; and a contact method. This might all pop-up
automatically if a known endangered or threatened species is reported.
The pop-up would also encourage a more detailed narrative. (and, hmmm,
I'm just poking around to see if there's a "profile" page to fill in
-- sez you're working on it. Great! Yes, that at a minimum would help
others judge accuracy).
This is turning into a great site, and there's got to be some clever
balance between encouraging people to report what they see (or think
they do) and being able to somehow rate the accuracy of those
sightings for the scientific community to use.
The conversion among different XY projections is pretty gnarly. TOPO!
will do it by hand -- I think you can enter whatever XY you have, then
switch to decimal degrees and read that (also klunky). If you're
really nerdy, ArcGIS with the National Park Service AlaskaPak now has
a feature where you can click on a point and it's automatically filled
into a clipboard to be copied wherever in whatever projection you
want.
George
On Apr 8, 7:45 pm, Ken-ichi <kenichi.u...@gmail.com> wrote:
> Hey April,
>
> Thanks for signing up! Here are some replies to your excellent questions:
>
> > 1. I have the GPS coordinates for some things that I saw and would
> > like to input that instead of searching for their location on the map.
>
> When you make a new observation, you can click the "Add additional
> data" link and you will see latitude and longitude fields. You can
> enter GPS coordinates there, but they'll have to be in decimal degrees
> (negative values for degrees longitude west). If you have GPS
> coordinates in other formats like degrees/minutes/seconds or the
> dreaded degrees/decimal minutes, you'll need to convert them using
> something like this:
>
Agrin Nathan
One thing that I think we should remember is that we want to provide
the community with a way to catalog, sort and manage their individual
observations outside of the scope off the larger community. That's
easy to forget because once the data is combined there are such
profound and interesting effects.
Managing data quality of the larger community has been a frequent
concern of every person we have interviewed, but ultimately we can
only do so much to allow for people to manage their data in the
context of everone else's. I suppose what I'm trying to say is that
I'm opposed to stringent data quality measures that potentially
exclude the interesting, bizarre and rare things people honestly see
in the name of commumity accuracy.
Now that's not to say we aren't concerned about quality, and in the
next few weeks you should be seeing some updates that address this
concern. One in the way that users can suggest identifications of
observations and another in how users can flag particular observations
with certain adjectives.
So bring on the wolverines, and keep the comments and suggestions
coming.
Thanks again,
Nate
Lirpa
can still be wrong with their ID's.
Here's the website for the California Department of Health
Services.The public reports dead birds and the CDHS uses this info to
track the spread of WNV. The concept is a little bit similar to what
you guys are doing.
http://www.westnile.ca.gov/bird_report_id.htm
I also saw a different site that used different colored dots to
represent confirmed vs unconfirmed sightings... but I can't find it
now. Instead of that website I found a different one that compiled
cougar sighting maps. This one uses a lot of data.
http://www.uwsp.edu/wildlife/carnivore/Mountain%20Lion%20Maps_files/Mountain%20Lion%20Maps_copy(1).htm
I like the rare mammal observation card. It's a good way to screen
observations.
Just a random question, can people input information on the evidence
of an animal too? Like if they found tracks or scat?
-April
George
ranger for the NPS in Sequoia Kings Parks. We record our wildlife
sightings and send them to the biologist. Sequoia has, I think, 30,000
+ records going back to the late 1800s. The early data, especially,
was gathered by reading old Superintendent's reports -- and the origin
of the record is noted in the database. So there'd be some error
definitely possible on individual records -- especially in the early
ones -- but when you start getting numbers like that, you can see
stuff start fitting into patterns (or not) and become more reliable.
I'd hope for the same here. The other thing about this database is
certain names start emerging with a lot of observations, and you tend
to start trusting those consistently.
I've done a number of observations -- almost all bear -- based on
scat. I've never actually checked to see if this is kosher, but I'm
pretty confident of coyote, bear, marmot to report based on that. I'd
do the same here, but note what it's based on.
I'm not as confident about tracks, though I've done that, especially
where I think it important -- a raccoon above 9,000' in Kings Canyon.
The occasional coyote in very good tracking conditions -- I ran into a
wildlife biologist who was writing a paper who said there were no
coyote sightings above something like 8,000 feet in winter. I told him
-- based on tracks, hair and scat, that I fairly often see sign above
10,000 and he seemed to accept that.
Which is all to say that I'd support this site accepting tracks or
scat-based observations but being sure to note the evidence. I'm sure
you guys are coming up with a sort of "how to/guidelines" for
observations?
Ken-ichi & Nate??
Thanks for the dead bird recording site. Didn't know that existed.
Also, somewhere (I'll look for it later) is a site that asks for the
flowering dates of plants. I think it maps them or somehow makes the
data available. The same could be done for when a bird species first
shows up (probably something for that already...).
George
On Apr 9, 6:45 pm, Lirpa <ap2il_y...@yahoo.com> wrote:
> I like the optional data rating thing... but even experienced people
> can still be wrong with their ID's.
>
> Here's the website for the California Department of Health
> Services.The public reports dead birds and the CDHS uses this info to
> track the spread of WNV. The concept is a little bit similar to what
> I also saw a different site that used different colored dots to
> represent confirmed vs unconfirmed sightings... but I can't find it
> now. Instead of that website I found a different one that compiled
Ken-ichi
One of our test participants mentioned that she would like to record
occurrences of tracks and scat, so it's interesting to hear that
echoed. They're certainly legitimate and useful observations, and if
people want them, I'm all for them. Currently, there's nothing
stopping you from making these kinds of observations, so I guess the
real question is would you trust those observations less? Would you
want to filter them out from the larger pool of observations? Do you
see a need to distinguish them?
And thanks for all the links! Another curious site along the lines of
what you were talking about, George, is the British site Nature's
Calendar. They use observational data to make interesting animated
maps of season phenomena. Check this out:
http://www.naturescalendar.org.uk/map/current.htm?rsid=164&reid=3&yr=2008&rs=S
I have a small collection of similar iNat-related links at
http://del.icio.us/kueda/biodiversity+informatics if you're
interested.
-Ken-ichi
George
a drop down box, number 1 to 5, with text indicating the confidence
level with examples:
1: Fleeting glimpse or based on scat or tracks or other sign (e.g.
scratching on tree -- bear)
2: Fair look, but not 100% sure because of either experience or
similarity to another species; or based on scat, tracks or other sign.
3: Good look. 90% certain; or based on experienced knowledge of scat,
tracks or other sign
4) Absolutely sure. Sufficient observation period to identify and not
confuse with similar species (similar species eliminated as
possibilities)
5) Absolutely sure. Photo provided.
Hmmmm. That's pretty rough, but you get the idea. Worth discussing???
George
On Apr 10, 10:13 pm, Ken-ichi <kenichi.u...@gmail.com> wrote:
> Hey all,
> One of our test participants mentioned that she would like to record
> occurrences of tracks and scat, so it's interesting to hear that
> echoed. They're certainly legitimate and useful observations, and if
> people want them, I'm all for them. Currently, there's nothing
> stopping you from making these kinds of observations, so I guess the
> real question is would you trust those observations less? Would you
> want to filter them out from the larger pool of observations? Do you
> see a need to distinguish them?
>
> And thanks for all the links! Another curious site along the lines of
> what you were talking about, George, is the British site Nature's
> Calendar. They use observational data to make interesting animated
> maps of season phenomena. Check this out:
>
>
> I have a small collection of similar iNat-related links athttp://del.icio.us/kueda/biodiversity+informaticsif you're
> interested.
>
> -Ken-ichi
>
> read more »