Utah Birds, Utah Birding, and Utah Birders. Promoting the sharing of information, and the conservation of habitat for birds in Utah and elsewhere. We are a group of people who want to share what we know, and create a positive birding experience in Utah.

BIRDERS BLOG

a blog by and for Utah Birders

NEW Historical Data in eBird

posted by Tim Avery at
on Friday, September 19, 2014 

After my eBird "quality" posts last week, which started as a commentary on making historical checklists more valuable, I laughed when I logged into eBird this week and noticed "historical" was now an option for entering a checklist!



Team eBird has basically stated that to make data even more concise they've changes a few things.  Historical would be used in place of incidental, or either traveling or stationary counts.  Here is what they have to say about historical data:

The value of entering historic data into the eBird database cannot be overstated. These records help build historic perspective in the database, and allow us to look farther back in time when conducting analyses. In these cases it's good to have as much supporting documentation available for the old records as possible, and be prepared to field questions from our regional editors about the validity of the data. All eBird records are treated the same way, going through our data-vetting procedures, and historic data are no different.

Historic data entry does have a few caveats though which you can read about here.

But there is more.  Team eBird apparently also wants all counts to be either stationary or traveling.  If you are birding, or a heavy portion of what you are doing is birding, they want you to use those count types. Here is a quote directly from an email that was forwarded to me about the change:

Our goal is to encourage people to use Stationary and Traveling counts. These are the two principal types of surveys that provide the most useful data (aside from some specific surveys we may have in the drop down menu for particular projects). When we have start time, duration, distance, we can do much more with the data particularly when coupled with a yes to the question about whether all species were detected.
After looking through a large amount of data, it became clear that there are two different ways that people were using the "old" Incidental protocol. One was for historical data entry where people were simply missing values but had made an attempt to survey an area for a reasonable amount of time. The other was for current checklists. Of current checklists, we found tens of thousands of records with one or two species where people said they were answering yes to the question, are you reporting all species. This is NOT the intent of that question and we needed to address it.
The question that I'll paraphrase as "Are you reporting all birds" has a temporal component.  In order to report "yes" to the question of are you reporting all species of birds, there must be a dimension of time, otherwise the answer doesn't make much sense. If someone spends 5 minutes surveying for birds or 10 hours, there will be very different number or birds detected. The same is true for distance. If we have no information on this, it doesn't help much to have an answer of Yes to this question.

You'll now see this message near the submit button on your checklist if you select incidental--and you can't change it to "complete checklist".  This means that even if you stop for 1 minute to bird a location--and you are really birding and only see 2 species, you would submit a stationary count as opposed to incidental. It's a great  way to improve data quality.


I talked with some friends recently about dozens of old field notes from when I did field work in Wyoming. 99% of those notebooks aren't in eBird--along with 10+ years of birding before that.  I have 1,000's of historical checklists just laying around that one day I want to enter in eBird--and now there is a count type specifically for them!  To that I say hooray! I will also have to start using the stationary and traveling more often when I make a quick stop while birding.  I guess those minutes are important data not to be dismissed as incidental after all!

Labels: , ,

Making Your eBird Data More Valuable PART DEUX

posted by Tim Avery at
on Thursday, September 4, 2014 

Here is my reply to ALL the replies form yesterday post...



Wow, what a great batch of comments, first I must commend Shyloh, Mike, and Kenny for publicly sharing their views.

The 11 anonymous posters had a varied mix of feelings and thoughts, and some obviously didn't want to share their names despite such strong convictions.  I want to respond to every comment so bare with me while I go through them...


Anonymous 1:

This seems like pretty blatant trolling--you say you don't care, but the time you took to post this with your feelings shows otherwise.  You say, "I know what I saw", which is fine, but if your sightings seem questionable, and have little details, their addition to the public data poll is a bit careless.  Like I mentioned, I am okay with people using eBird to just track their personal sightings for themselves--its a great tool for that.  However, for anyone that feels this way, I highly recommend going into preferences and checking the "hide my data" checkboxes for eBird Alerts, and Top 100 under the data privacy section.  If your sightings are just for your personal use as you say they are, there is no need to have them included in these very "gamified" areas of eBird as one other commenter noted.  My last criticism of your response was not having the time to enter a bunch of lists.  You have time to go birding, you have time to enter at least one list, and you have time to comment on this blog.  So time is on your side I would argue, and the little amount of time it takes to break 1 list into 3 or 4 seems well worth it this day in age.


Anonymous 2:

Great comment.  As I mentioned to Anonymous 1 this is the logical thing to do for others, and for yourself for a multitude of reasons.

Anonymous 3:

I see the first comment got you riled.  I have to say I disagree with your first point--eBird has many uses, and personal lists are a valid reason to use the program--its unfair to tell someone to use a notebook because they don't give a shit about science when the software is available for free and their data can be invalidated from the pool.  As to your second point, I do tend to agree, but if someone doesn't use it to help with the scientific effort, their data is so minuscule in the bigger picture, that in 150 years it really won't matter.  Those "Rare" birds they report, probably were mis-identifications, so their exclusion from the data pool is probably for the better.  And as to your 3rd point--haha, spot on my friend.



Anonymous 4:

I think in general in our local area--Utah, or even just northern Utah, the majority of eBird users are aware of what sightings are legitimate, and which are questionable.  I would be lying if I said I didn't brush off a number of alert species based of the observer.  In our community there are a handful of eBirders who don't so their sightings justice by taking photos of rarities they reports, or providing any meaningful information on the sightings for both eBird and their peers.  Before eBird it was much the same with the "RBA" or just on the listserv's. In birding credibility is earned not through documentation.  If a supposed great birder never documents his/her sightings, but constantly reports rarities, it begs to question the legitimacy of the sightings.  IF other give chase and can't track down the reports, it also leaves you to wonder.  But a photo, and re-chaseable birds create reputations and credibility amongst the community.  I expect every level of birder from beginner to expert to provide the same level of documentation in eBird for the sighting to really resonate with me.  This is completely aside of the scientific aspect of eBird, but more on the who is full of shit front.


Anonymous 5:

As I mentioned before, this is an opinion, and someone who just uses it for themselves is not breaking rules or being selfish---eBird has many uses.  I do believe people who use the applications should contribute more, but by no means do they have to, and by no means are they "doing it wrong".  But they should take care to remove themselves from alerts and listing.


Anonymous 6:

This is true--however an X can denote a count that is too high for a decent estimate.  Many birders don't feel comfortable, or can't estimate high counts of birds.  It sounds crazy but I have done this exercise with people to see how they estimate, and counts are wildly off sometimes.  If it wasn't an easy to count exact count, it may be best to leave it as X--this data is still important and useful.  I always try to put a number in, and in 99.9999999% of my individual sightings there is a count, whether exact or an estimate.   It definitely adds value even with estimations though!



Mike:

Woot an name and a face!  You are nuts, a nutty nut job, blah blah blah.  Just kidding of course, I just figured a name made it easy to call you out!  In general it seems like you feel how I feel.  I'm okay with people using the program how they see fit--I love when there are more details, and it lends credibility, and its better for the data pool.

Kenny:

You really are crazy... Crazy in a good way.  I think a lot of people miss the communal aspect of eBird--whether or not you want to be part of that community, you are if you report rare birds, and take part in the Top 100.  These things are far less important that high value checklists, but they have become a big part of eBird.  You said it best though, "We should all strive to be a better part of the birding community since who can understand the highs and lows and adventures of birding best than other birders?"  Providing better data inherently makes you a better part of the community--and our growing community here is proof of that!


Shyloh:

Where do I even start with you man... You're the dude.  You make the point right off the bat, "THERE IS NOTHING WRONG WITH SELFISHLY USING EBIRD!!!"  And Like I said, I agree--use it as you see fit, its there for many different reasons, so to each their own.

At the end of the day it really doesn't matter--like I told you, they're just birds man.  But I like to think that striving for a strong community, with birders that respect one another includes some great self checking, as well as community checking--which helps build a great data pool.  If someone reports rarities often, or even not so often, but rarely produces chase-able birds, or documents their sightings it effects not only the data pool, but chasers and listers--the community.  Regardless of how people use eBird, if you report a rarity on a checklist, you should expect some level of scrutiny from others, that's just human nature, and a big part of birding--and quality documentation and information helps eliminate some of that.

I would never tell anyone "this is how you should use eBird", but I sure as hell think there are ways we can all make our checklists more valuable not only for the community but for ourselves!  And I think that is a lot of what others are getting at too--just a tad more forceful.



Anonymous #7:

Shyloh, wanted to tell you "LOL"...

Anonymous #8:

Holy moly.  I will just say this, well played sir.  You hit the nail on the head.

Anonymous #9:

This is a great point with all others aside--as volunteers, having to wade through the muck, it sucks.  Whether a user cares about listing, or conservation, or whatever, a sighting that lacks documentation makes their lives harder.  I have talked to several moderators who routinely throw out specific users sightings because they are so riddled with mistakes, and absolutely no documentation that it's better for the data pool to just remove them.  I can see eBird eventually coming up with a way to block certain users from the data pool based of reviewer recommendations because of just that.

AND NOW MY FINAL THOUGHTS...

Let's just be real about our sightings for a second--on a personal level.  In the end on either side (community/data pool versus selfishness), it comes down to do you care about being right, or thinking your right?

I personally would rather be able to look at my sighting, have documentation, and be able to say I am 100% sure of that--and so is everyone else that sees it, than have a sighting with no documentation that begs the question, is this reliable data.  Simply putting a tick mark on a checklist for a rare birds shows a lack of understanding why that bird is rare/flagged, and the importance of documentation for data quality, especially when its not just a checklist for keeping a life list, but one used for building a data set for a location.  In the end the reviewer can simply invalidate the sighting, but then it seems counter intuitive to the whole process.  I don't think its asking to much for even a crappy photo--almost everyone carries a digital camera with them, and rarely are birds so flighty that even a distant, blurry, out of focus, or partially hidden bird can't be photo'd.  It seems so simple, and useful to the whole idea of building the data pool.

One solution to this is creating user groups in eBird for different types of players in the "game".  For the selfish, I use it to keep my life list, there could be a user type where all data is separate from the community pool.  Then there could be scientific users, and typical users, so on and so forth.  You could even go to a pay model which is a whole other blog post...

No matter what reason you watch birds, if you are using eBird you have made yourself part of that community, and we're glad to have, and I hope we all can work to make the community stronger in the long run! At the end of the day every birder is different, and uses eBird for many different reasons--I just hope that everyone realizes that we can all strive to make things better!

Labels: , , , ,

Making Your eBird Data More Valuable

posted by Tim Avery at
on Wednesday, September 3, 2014 

I recently posted a response to a thread on UBIRD about historical checklists and incidental counts in eBird.  There were 3 main topics I talked about.

  1. Adding valuable sighting information to historical vagrants/rarities
  2. Splitting larger generic checklists into smaller location specific checklists
  3. Removing possibly bad data or in particular mis-identifications
The email didn't go in to a ton of detail on these topics, just my brief overview of why I felt it was important.  But I wanted to expand on that, and give a few real world examples of the issues, and what I've been doing to try and practice what I preach.  I've had a few interesting conversations about the subject, and gotten different perspectives and personal opinions on the matter.  It brought up a couple things that I thought were also worth talking about.

I talked to several people who mentioned their #1 reason for using eBird was to keep track of their sightings for themselves above all else.  The data after all was being provided free of charge by the birders/observers, and that in itself was valuable despite the quality of the data.  It would be a lie if I didn't say that was also WHY I initially started using eBird back in 2004. It was an easy way to track my own sightings, I was after all filling up several notebooks a year with lists and this was an easier way to keep those lists, and manage them.  For this reason I was quite cavalier with my locations.  I often did counts for entire days for a county, or along 100 mile section of road.  My lists were less for science than my personal needs.

There are in fact a lot of eBird users who ONLY use eBird for this very reason, and that's perfectly acceptable--it's a great tool for that--but it's also so much more.

Now I also talked to one person who mentioned that sometimes they just put in a list for a whole weekend because they wanted to keep track of the birds they'd seen, but they'd visited 10 places, seen 150 birds, and didn't want to take the time to go enter individual checklists for every place they went.  That's another one I could relate to--I've done exactly the same.  Creating and submitting checklists can be time consuming, and for some they just don't have the time to get granular. Some data is better than no data after all.

Cape May Warbler from Lytle Ranch... Less than 5 ever reported in Utah.

And yet one other person who kept very detailed lists of species seen never added details, photos, or any information to their lists.  In their opinion they were already providing high quality data, and felt that the need to add more details was benign.  In the long run there will be so much data that these details won't matter--the pictures, notes, etc. And again, I have done the same, I get it, and in some cases I whole-heartedly agree.  I have heard the argument that in 150 years someone might be looking back at some data and see a report of a Cape May Warbler in Utah with no associated data.  The report was validated but without any supporting data because I was experienced with that species.  But how would those people know who I am?  Honestly, I hope I leave some type of impact on birding in Utah in my life time, whether small or big--I would hope with a little digging someone can figure out who I was--especially in this digital age.  But I digress, I also see the flip side, and that's where this whole blog post got started.  Let's talk about those bullet points.

Adding valuable sighting information to historical vagrants/rarities.

Over the years I have slowly become more and more methodical with my checklists.  This means being more granular with locations, and more detailed with subspecies, and adding specific details about flagged sightings.  This is a personal choice, to try and make my checklists not only more valuable to researches, and those using the data for scientific purposes; but also, for other birders to be able to glean them for data, and for me personally to be able to look back and see details about a sighting all in one place.  As of the writing of this I have submitted 3226 eBird checklists for Utah.  As that number grows it becomes harder and harder to look at specific checklists with flagged sightings and remember all the details. So thus started the drawn out process of going back and adding those details.

Utah's 1st documented Bay-breasted Warbler

I started with my taxonomic life list for Utah and started working my way down.  I specifically wanted to concentrate on adding details to species with less than 25 reports ever for Utah--with more details being added to the rarest species--those typically with less than 5 reports ever.

Utah's 1st Yellow-bellied Flycatcher 

In cases where I had photos, videos, or recordings I embedded those in.  Photos linked to my website, videos from Youtube, and recordings from Xeno-canto if available.  If I felt the photos were sufficient for the report I often kept the notes short.  On some species I added a small blurb, and on others complete paragraphs of texts that I cut out of emails I pulled from the original reports, or from Records Committee Reports from some of my oldest sightings.  I didn't remember all the details from many of these reports so going back and searching in UBIRD, and other listserv archives for my original posts was really helpful.  It also saved me from writing in details that might not be as accurate otherwise.

Utah's 1st Pine Warbler

I soon realized I had hundreds of such checklists, so it was somewhat tedious--but the end results were lots of photos added into the data pool, and some detailed (and not so detailed) write-ups of species that otherwise were originally note-less.   In all so far I've added details for 36 species, on 52 checklists.  Some of the more notable were details for eBird Utah's first Western Gull, Pine Warbler, Eastern Meadowlark, Sharp-tailed Sandpiper, Cape May Warbler, Yellow-bellied Flycatcher, Gilded Flicker, Bay-breasted Warbler, Black-billed Cuckoo, and Little Gull..  Despite the time that went into it, it was fairly easy to do, and well worth it.

Splitting larger generic checklists into smaller location specific checklists

I recently flipped through a field notebook from one of my summers working in Wyoming during college.  I had 1 or 2 checklists for most days.  One list was the survey I ran that morning, and the other was the rest of the birding that day.  In some cases I drove over 100 miles on these days and went through every typical habitat type found in western Wyoming.  In eBird these lists would look ridiculous--but none-the-less I have a few.  Luckily most of my college years checklists from Wyoming are not in eBird yet. But I have dozens form the early years of eBird that I did enter like this.  Some are an entire day count for 16 hours at a lake in the Uinta's.  The checklist covers shrubsteppe, aspen, conifer, and mountain reservoir habitats though.  I probably traveled 10 miles during the day and the list could be broken down more.  These are hard though, because some of the birds cross habitats, and I don't remember what was in which necessarily.  For these it's really best to just leave as is--its 1 checklist out of 1,000's, in the bigger picture that data is "close enough" that it won't look too out of place--and if it is the reviewer can invalidate the list.

Blue Grosbeak from 12,000' in the Uintas??? I think not.

But what about a 40 mile drive from Vernal to Hacking Lake in the Uinta's?  It looks odd when a checklist has both Blue Grosbeak and Pine Grosbeak together.  Or Red Crossbill and Lesser Goldfinch.  Or Dark-eyed Junco and Marsh Wren (in August).  During the early years I did this often--one checklist for the entire length of a road.  The best recommendation is going through and removing the species that don't fit the greater picture, and breaking them out into a separate list at a better location for them.  Or break it down in to as many lists as you feel comfortable doing.  In this case I broke 1 list into 5.  Vernal, the lower section of Sagebrush, Aspen Forests, Conifer Forests, and finally Hacking Lake.  It's still not perfect, but it fixes one bad list.

Pine Grosbeak should never be on the same list as Blue Grosbeak in Utah...

One of the most common offenders of this is the "Antelope Island State Park" checklist submission.  There are numerous hot spot points for the park, but many folks do one list for 5 or 6 visited spots.  This means you have Snowy Plovers and Godwits on the island.   Others mistake the causeway spot for the whole park and then you end up with Warblers, Tanagers, and Orioles on the causeway.  I tend to do a list for the causeway and each major location I go to on the island.  Sometimes I leave with 7 checklists--but its super granular and useful.  This is one thing I have made a better effort at in recent years and try to be as specific as makes sense for the places I'm going.  With apps like BirdLog it now makes it easier to avoid this, allowing you to quickly create a new list on the fly while in the field as you go from habitat to habitat, or location to location.

Removing possibly bad data or in particular mis-identifications

This is one of the hardest for most people to do--myself included.  Going back and looking at lists and saying, that's not right, or I'm just not sure anymore, can be difficult for most birders.  I went back and removed a number of my old Short-billed Dowitcher reports and changed them to Dowitcher sp.  I did leave ones that had photo evidence, but ones without were mostly changed.  I went back and took Least Bittern off an Ouray list I had form my younger years.  At the time others were reporting them, and I thought I saw one.  But looking back and having seen that species since it was a no brainer to remove it.  In some cases if it doesn't seem right, its safer to just change it outright.  Leave a note, change it to a spuh, with details about why you changed it.  But if you aren't sure this is a good way to clean up old lists.

Dowitcher sp? Or one of each?

There are some that I probably should change, but I can't bring myself to do it.  Sometimes you believe so much in a sighting you had, despite lack of evidence, experience, etc, you just stick with it.  You very likely are correct with your belief, and in many cases the data has already been invalidated in eBird, so for your personal use what does it matter?  It's all about finding the right balance for how you use eBird.

Whether your a hard-core eBirder, a casual user, or just someone who keeps basic lists via the awesome resource, there are things we can all do to make our lists more useful--for everyone.  And I'm sure there will be more ways in the future, and things will change, and improve.  eBird has come along way in the last 10 years, and so has the quality and amount of data.

How do you personally make your checklists more useful?  Share below, and help others improve the data they're submitting!

Labels: , , , ,




UTAH BIRDERS FLICKR POOL


    SEE MORE AND SHARE ON FLICKR