eBird duplications

[John Loch]

Hi,  I've encountered what I find a significant barrier to doing my job as a sub-editor for 2 sections of NS Birds: duplication of records and therefore inflation of sightings.

The following email (Jan. 27, 2017) outlines the problem (it's an email I sent to Dave C, Ian M and Keith).  Does anyone have a workaround to this issue?  I would note it's an issue for editors, etc. and not for birders, per se because their records are fine, it's compilations thereof where the problems start.  One example, geographically, is BPI where I'm getting 1-5 duplicates for most records submitted.  That is adding up to literally thousands of duplicates or more.  The data cited below is an early estimate as I've now gone through all of the relevant species and there are a lot more duplicates.  Note also that Keith replied that the only solution he sees is to use the "share checklist" option.

"In going through my two groups (Tubes to Corms and Skuas to Terns), I have discovered that there are a lot of duplicate records and therefore duplicate sightings numbers.

I spent a lot of time (>>50 hrs) trying to identify and delete all duplicate counts.

Finding these duplicates in 56,594 records is not simple, obviously, unless one has a macro or some other multiple sorting and flagging/identifying duplicates algorithm.  I used multiple sorts in Excel and that helps a lot but one still have to make a lot of decisions on individual records.

I have written very recently to eBird to see if they can deal with this at “the request to eBird” stage.  I haven’t yet received a reply.

My 2016 Fall data base came from eBird via Keith Lowe; the earlier “all years from 2006-2015” data set was requested by me directly from eBird a year ago. Both sets have the same duplicates problem.

I wanted this longer term data base in order to situate the  current year fall results with the longer term (10+ years) trends (hopefully some day!) and averages.

As you can see from the Table below, this duplicates issue is not insignificant for the Skuas to Terns groups for certain - 29% or 430,037 duplicate sightings -  and, perhaps, marginally (50,662 or 8.8%) for the Tubes to Corms group as well.

I am not sure who should fix this but I believe it applies to all NS Birds’ data sets whether directly from eBird or through Keith.

There may be others that should see this email as well? (e.g. Dave mentioned Dominique last night after the NSBS meeting when I was speaking to him about this matter …)

 

John"

 

Tubes to Corms	# dup records	total records	% dup records	# dup sightings	total sightings	% dup sightings #s
2016	78	3,920	2.0%	1,868	111,232	1.7%
2006-2015	542	21,741	2.5%	48,794	462,413	10.6%
totals 2006-16	620	25,661	2.4%	50,662	573,645	8.8%
Skua to Terns	# dup records	total records	% dup records	# dup sightings	total sightings	% dup sightings #s
2016	85	9,422	0.9%	13,361	337,724	4.0%
2006-2015	5,030	21,511	23.6%	416,676	1,122,975	37.1%
totals 2006-16	5,115	30,933	16.5%	430,037	1,460,699	29.4%
both Groups	# dup records	total records	% dup records	# dup  sightings	total sightings	% dup sightings #s
2016	163	13,342	1.2%	15,229	448,956	3.4%
2006-2015	5,572	43,252	12.9%	465,470	1,585,388	29.4%
totals 2006-16	5,735	56,594	10.1%	480,699	2,034,344	23.6%

[Phil Taylor]

Hi John, 

In ebird, a record is an observation of a single bird at a given location and time (within the context of a 'checklist' which provides additional data about effort (time of day, duration, number of species, area searched)

So when multiple people see the same bird, those are different records, which I think is what you are calling duplicates. But they aren't really ... it just isn't the information that you want.

I think you want some idea of the total number of individuals of a given species that were likely present in a given location over a given time period. To do that, you need to aggregate to some other level, by either taking a maximum, or a sum of a set of maxima, or something else that captures what you want. Regardless, you probably need to make some assumptions (and although often our assumptions about there being a single bird of a given species present in some location are likely correct, they are still assumptions) when you decide how to aggregate. 

 

I can imagine that some of these summary tools might already be available, or at least are being considered ... in the interim, my students and I (David? Lucas?) could probably build them and provide some of the summaries you're looking for.  If that was of interest ...

Hope this helps, forgive me if this is all obvious to you and I'm misunderstanding your question ... 

Phil.

--
You received this message because you are subscribed to the Google Groups "NS eBird discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsub...@googlegroups.com.
To post to this group, send email to ns-ebird-discussions@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/3c50d543-b18a-4c4a-aa1b-2321b019e476%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Phil Taylor
Bird Studies Canada Chair in Ornithology
Acadia University, Wolfville, NS.

[John Loch]

Hi Dr. Taylor, I don’t believe that we’ve actually met per se although I did attend a NSBS meeting where you gave a presentation!

 

I think it’s best if I start at the beginning to try to ensure there’s a common understanding of the problem and how it became one:

 

·         As you know, NSBS has a ‘popular, quasi-scientific birding journal’ (my words) that is published seasonally, that is 4 times a year, via “NS Birds”.

·         Without getting into the history, suffice to say that I am the current sub-editor for both the Tubes to Cormorants and the Skuas to Terns sections and am responsible to write up these species groups’ sections four times a year.

·         My understanding of the purpose of NS Birds (in terms of my own responsibility) was to provide up to date information on sightings of birds for the groups I was responsible for.

·         I soon learned that the primary source for sightings data was eBird and that Keith Lowe provided relevant data bases in Excel from eBird for each sub-editor.

·         There are other secondary sources of data that one also checks (NS Nature List Server, Facebook, NS-RBA) but these provide miniscule amounts of sightings relative to eBird.

·         (In my particular situation where I’m covering seabirds, I also arranged to obtain the relevant (to NS waters) data sets from Carina Gjerdrum of CWS from their off-shore surveys via ECSAS but this isn’t relevant to the matter-at-hand.)

·         After getting my feet wet with the data sets and the nature of the task-at-hand, I became interested in long term data on sightings in order, initially, to situate this season’s sightings with previous ones for that particular season.  That is I wanted to get the sightings data from eBird for all falls since 2006 so I could show and discuss how there have been inter-annual changes in sightings numbers as well as comparing the current season’s sightings with long-term, decadal, averages.

·         In order to obtain that long term eBird data, I requested it directly from eBird and it was quickly provided.  Again, that data set was & is in Excel.

·         With a  couple of seasons behind me, I began to realize there were frequent  situations where there was more than one “record” for many species with the same sightings data (count, location, date). Here is an example from Fall 2015, noting that this is but a small subset (53 records [with counts of >= 100 birds] of 275 records) for one species for one month (of a four month season) for one (of eleven) years:

 

COMMON NAME	OBSERVATION DATE	LOCALITY	LATITUDE	LONGITUDE	OBSERVATION COUNT	Count Adjusted for duplicates	TRIP COMMENTS	SPECIES COMMENTS	LAST NAME	FIRST NAME	OBSERVER ID
Northern Gannet	30-Oct-15	Bon Portage Island	43.46937	-65.7513	2000	2,000			Bartels	Avery	obsr108813
Northern Gannet	30-Oct-15	Bon Portage Island	43.46937	-65.7513	2000	0			Bell	David M.	obsr121486
Northern Gannet	30-Oct-15	Bon Portage Island	43.46937	-65.7513	2000	0			Brown	Morgan	obsr346534
Northern Gannet	30-Oct-15	Bon Portage Island	43.46937	-65.7513	2000	0			Cormier	Dominic	obsr206296
Northern Gannet	17-Oct-15	Cape Sable Island--Daniel's Head	43.44158	-65.5902	600	600		Approx. We had a big southerly rainstorm, I watched as soon as it stopped, counting 87 past in one minute! The passage slowed after five minutes or so.	Dennis	Mark	obsr338688
Northern Gannet	03-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	500	500			Bell	David M.	obsr121486
Northern Gannet	03-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	500	0			Cormier	Dominic	obsr206296
Northern Gannet	20-Oct-15	Bon Portage Island	43.46937	-65.7513	500	500			Bell	David M.	obsr121486
Northern Gannet	20-Oct-15	Bon Portage Island	43.46937	-65.7513	500	0			Cormier	Dominic	obsr206296
Northern Gannet	05-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	400	400			Bell	David M.	obsr121486
Northern Gannet	05-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	400	0			Cormier	Dominic	obsr206296
Northern Gannet	29-Oct-15	Baccaro Peninsula--Baccaro Point	43.45018	-65.471	400	400	Big storm, E-S/E 6-7,  heavy rain at times.	Approx, all south.	Dennis	Mark	obsr338688
Northern Gannet	07-Oct-15	Bon Portage Island	43.46937	-65.7513	350	350			Bell	David M.	obsr121486
Northern Gannet	04-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	300	300			Bell	David M.	obsr121486
Northern Gannet	04-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	300	0			Cormier	Dominic	obsr206296
Northern Gannet	10-Oct-15	Bon Portage Island	43.46937	-65.7513	300	300			Bell	David M.	obsr121486
Northern Gannet	10-Oct-15	Bon Portage Island	43.46937	-65.7513	300	0			Brown	Taylor	obsr429936
Northern Gannet	10-Oct-15	Bon Portage Island	43.46937	-65.7513	300	0			Cormier	Dominic	obsr206296
Northern Gannet	03-Oct-15	Baccaro Peninsula--Baccaro Point	43.45018	-65.471	260	260			Dennis	Mark	obsr338688
Northern Gannet	23-Oct-15	Bon Portage Island	43.46937	-65.7513	250	250			Bell	David M.	obsr121486
Northern Gannet	23-Oct-15	Bon Portage Island	43.46937	-65.7513	250	0			Cormier	Dominic	obsr206296
Northern Gannet	01-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	230	230			Bell	David M.	obsr121486
Northern Gannet	01-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	230	0			Cormier	Dominic	obsr206296
Northern Gannet	03-Oct-15	Brier Island--North Point	44.28583	-66.3425	200	200			Stern	Richard	obsr16727
Northern Gannet	03-Oct-15	Bon Portage Island	43.46937	-65.7513	200	200			Observatory	Atlantic Bird	obsr280927
Northern Gannet	03-Oct-15	Brier Island--North Point	44.28583	-66.3425	200	0			Walker	Jake	obsr44100
Northern Gannet	03-Oct-15	Brier Island--North Point	44.28583	-66.3425	200	0			Whitman	Rick	obsr309800
Northern Gannet	31-Oct-15	Bon Portage Island	43.46937	-65.7513	200	200			Bartels	Avery	obsr108813
Northern Gannet	31-Oct-15	Bon Portage Island	43.46937	-65.7513	200	0			Bell	David M.	obsr121486
Northern Gannet	31-Oct-15	Bon Portage Island	43.46937	-65.7513	200	0			Brown	Morgan	obsr346534
Northern Gannet	31-Oct-15	Bon Portage Island	43.46937	-65.7513	200	0			Cormier	Dominic	obsr206296
Northern Gannet	06-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	150	150			Bell	David M.	obsr121486
Northern Gannet	06-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	150	0			Cormier	Dominic	obsr206296
Northern Gannet	09-Oct-15	Bon Portage Island	43.46937	-65.7513	150	150			Bell	David M.	obsr121486
Northern Gannet	09-Oct-15	Bon Portage Island	43.46937	-65.7513	150	0			Brown	Taylor	obsr429936
Northern Gannet	09-Oct-15	Bon Portage Island	43.46937	-65.7513	150	0			Cormier	Dominic	obsr206296
Northern Gannet	16-Oct-15	Waterside Beach Provincial Park	45.75898	-62.7765	150	150	I spent the bulk of the time sitting and watching birds on the water		Scanlan	Peggy	obsr428628
Northern Gannet	20-Oct-15	Cape Sable Island--Daniel's Head	43.44158	-65.5902	150	150		Moving south.	Dennis	Mark	obsr338688
Northern Gannet	30-Oct-15	Cape Sable Island--The Hawk	43.41122	-65.6182	150	150			Dennis	Mark	obsr338688
Northern Gannet	30-Oct-15	Cape Sable Island--Daniel's Head	43.44158	-65.5902	150	150			Dennis	Mark	obsr338688
Northern Gannet	28-Oct-15	Bon Portage Island	43.46937	-65.7513	140	140			Bartels	Avery	obsr108813
Northern Gannet	28-Oct-15	Bon Portage Island	43.46937	-65.7513	140	0			Bell	David M.	obsr121486
Northern Gannet	28-Oct-15	Bon Portage Island	43.46937	-65.7513	140	0			Brown	Morgan	obsr346534
Northern Gannet	28-Oct-15	Bon Portage Island	43.46937	-65.7513	140	0			Cormier	Dominic	obsr206296
Northern Gannet	23-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	120	120			Bartels	Avery	obsr108813
Northern Gannet	15-Oct-15	Outer Tuskets--Seal Island	43.40956	-66.0146	119	119	MJB saw a dickcissel <br />Submitted from BirdLog NA for iOS, version 1.5.2		Bartels	Avery	obsr108813
Northern Gannet	04-Oct-15	Cape Sable Island--Daniel's Head	43.44158	-65.5902	100	100			Dennis	Mark	obsr338688
Northern Gannet	07-Oct-15	CA--NS--Yarmouth--boat to Seal Island	43.43733	-65.8871	100	100			Bell	David M.	obsr121486
Northern Gannet	07-Oct-15	CA--NS--Yarmouth--boat to Seal Island	43.43733	-65.8871	100	0			Cormier	Dominic	obsr206296
Northern Gannet	29-Oct-15	Bon Portage Island	43.46937	-65.7513	100	100			Bartels	Avery	obsr108813
Northern Gannet	29-Oct-15	Bon Portage Island	43.46937	-65.7513	100	0			Bell	David M.	obsr121486
Northern Gannet	29-Oct-15	Bon Portage Island	43.46937	-65.7513	100	0			Brown	Morgan	obsr346534
Northern Gannet	29-Oct-15	Bon Portage Island	43.46937	-65.7513	100	0			Cormier	Dominic	obsr206296

 

·         There are several things here that I want to flag:

o   “record” in the context of someone reviewing data received from eBird is a single row in  an Excel data set that contains a count of multiple sightings of the same species – this is not what you defined as a “record”, but I submit you were using that word as a birder entering data on an eBird virtual checklist, not as a reviewer of a data set received from eBird which has been assembled in a certain way requiring a “record” to be something quite different;

o   By using the multiple sort feature in Excel, one can organize the data set to quickly assemble it in a way that  you can manually note and flag apparent duplications of “records” (sensu John Loch) – see cells coloured in orange;

o   Using just the example above for one species (N Gannet) for one month (October) for one year (2015), there were 82 records of 11,800 gannet sightings; therefore  ~50% of the total number of gannet sightings (~ 23,000) for October 2015 were duplicates; Ian refers to this as “inflation of the sightings” due to the “duplication of the records”.

o   Given that there are 52,000 such records in the 2006-2016 data base, it was NOT a trivial task  to flag such duplicates

o   Given the magnitude and consequence of the “duplicates issue”, I felt it was necessary to do the manual sort and flagging but, at the same time, was and am looking for a more efficient way of doing this, given I literally spent easily 50 hours just on this duplicates matter.

 

·         As matters currently stand, Keith Lowe has reviewed this issue and he feels that the ideal solution would be for birders using eBird to share their sightings with all others who they were birding with at that time.  However, I doubt that this is practically possible for all kinds of reasons.

·         As such, I have asked eBird to provide a “macro” or algorithm or some sorting process within Excel to further minimize the time to flag duplicate records.  I suspect that this is the more pragmatic solution and believe that it’s possible to do this with Excel and its current sorting function but I need to spend some time on that and it’s not now, given I’m 2 weeks late in submitting my reports!

·         Finally, I’ve taken the liberty to comment on your email below using this backgrounder.

 

John Loch

 

 

 

 

 

From: ns-ebird-d...@googlegroups.com [mailto:ns-ebird-d...@googlegroups.com] On Behalf Of Phil Taylor
Sent: February 15, 2017 9:49 PM
To: John Loch
Cc: NS eBird discussions
Subject: Re: [NS-ebird-discussions] eBird duplications

 

Hi John, 

 

In ebird, a record is an observation of a single bird at a given location and time (within the context of a 'checklist' which provides additional data about effort (time of day, duration, number of species, area searched)  As indicated above, this is not my definition of a record nor is it eBird’s or any else who uses the databases provided by eBird.  This semantic issue can cause considerable confusion regarding the term “duplicates”, the heart of the problem, imho.

 

So when multiple people see the same bird, those are different records, which I think is what you are calling duplicates. But they aren't really ... it just isn't the information that you want. This is incorrect from the perspective of someone like me using eBird data bases to write up NS Bird sections where the number of sightings needs to be accurate and therefore duplicates should not be included in the sightings’ numbers.  Perhaps a good example is the Brown Booby seen at the Canso Causeway last fall – eBird shows 11 separate records by 11 people; so, eBird shows 11 sightings but there were 10 duplicates and only one legitimate sighting from a data synthesis perspective.

 

I think you want some idea of the total number of individuals of a given species that were likely present in a given location over a given time period. To do that, you need to aggregate to some other level, by either taking a maximum, or a sum of a set of maxima, or something else that captures what you want. Regardless, you probably need to make some assumptions (and although often our assumptions about there being a single bird of a given species present in some location are likely correct, they are still assumptions) when you decide how to aggregate. Not sure just what you mean here but I suspect that your “aggregation” is synonymous with my definition of a “record”!  Regardless, I think the solution lies in using Excel in a more complex way than I’ve currently been doing.

 

I can imagine that some of these summary tools might already be available, or at least are being considered ... in the interim, my students and I (David? Lucas?) could probably build them and provide some of the summaries you're looking for.  If that was of interest ... IMHO, the tools are available and they lie in Excel.  If your students have that capability and time, I’d much appreciate their help; however, I don’t want to get involved in non-Excel solutions.

 

Hope this helps, forgive me if this is all obvious to you and I'm misunderstanding your question ...   Thanks very much for your interest, Dr. Taylor, and I hope I’ve provided more clarity on this matter.

 

Phil.

 

--

To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discuss...@googlegroups.com.
To post to this group, send email to ns-ebird-d...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/3c50d543-b18a-4c4a-aa1b-2321b019e476%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

 

--

Phil Taylor
Bird Studies Canada Chair in Ornithology
Acadia University, Wolfville, NS.

--

You received this message because you are subscribed to the Google Groups "NS eBird discussions" group.

To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discuss...@googlegroups.com.
To post to this group, send email to ns-ebird-d...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/CAA1OBY%2BCZGdgmNwbQ7CgOsrUiCm7%2B9HG%3DbF8dtFq0fk9JiLXvw%40mail.gmail.com.

[Keith Lowe]

Further to the discussion on aggregating eBird entries.

First I want to note that when I said encouraging everyone to use shared checklists would solve most of the problem, I was not responding to John’s email here with that table of data. I was responding to some specific examples John sent me which were mostly duplicates of Brier Island pelagic trip last September. In the case of one-off pelagic trips, sharing checklists would solve the entire problem.

As for the problems in replicated entries in general, I discussed these same issues with Dave and Ian when I first considered processing eBird data for NS Birds. In the end it was decided I would aggregate by shared checklists only when I distribute the data for NS Birds. I group shared checklists into one record that takes the maximum count, the earliest start time, the longest durations, the longest trip comment, the longest species comment and appends all observers into one record. I also indicate if an observation was the first of year or first of season for that species. This was an easy step because eBird provides a group_identifier field for determining shared checklists.

But to take it beyond aggregating shared checklists  opens a hornets' nest of problems. Aggregating by public hotspots would be the next easiest step but aggregating data from personal hotspots really becomes difficult that would get into defining geocoordinate ranges. Also the definition of a "duplicate" is open for interpretation as well, such as duplicates per day, week, month season. For example, an observer notes 100 eider off Hartlen Point in the morning and another sees a similar number in the afternoon. These could be aggregated by day by public hotspot but what if one observer saw them on Monday and the other saw them on Tuesday. There is no biological difference between these two sets of observations so what is a duplicate?

To me, NS Birds does not claim nor imply that observations of common species are unique. If we wanted that we would far more  accurate just to report on the sum of the high counts of hotspots for a season but I would personally prefer it not be taken to that level for NS Birds.

But uncommon or less common species are treated differently by the NS Birds section writers. For instance it would be useless to report on the sum of Roseate Tern off Pond Rd, Pubnico. In this instance the high count per season would be the best number derived from data. Though a better way would be to get the number of pairs from Ted D’Eon. Similar story for puffin at Bird Islands or Pearl Island.

One of us should craft a post sometime to encourage more sharing of checklists though that is already trending quite well. It is especially important to share one checklist on a pelagic trip though.

It wouldn’t be too difficult to start aggregating by public hotspot but the question of by day, week, month or season would need to be addressed. But it may not be wise to put any effort into changes at this point until we see what changes might come about from eBird adding polygons to the data.  

[Keith Lowe]

I was perplexed at all the duplicate entries in your second  table John because it obvious many of them are shared checklists so they should be been aggregated into one record. I ran the section query for the same period but did not get the duplicates. I assume this data is something you did on your own and not from me?

Keith

[John Loch]

I believe that I have used your data sets since Fall 2015 but the earlier years (2006-2015) were obtained directly from eBird.

I'll check again though, Keith.

John 

[Phil Taylor]

Hi John, 

I understand the issue as you present it ... it is more or less what I indicated. 

I was a bit lazy in my definition of 'record' ... I should have said 'bird species' ... which is in fact what one of your lines is up above. 

"In ebird, a record is an observation of a single bird [SPECIES] at a given location and time (

within the context of a 'checklist' which provides additional data about effort (time of day, duration, number of species, area searched)"

... and that observation either has a count associated with it (0 or above) or an X (present). 

So, lots of people can be staring at the same bird at the same time, and those are all different (and valid) records. 

The issues that you describe regarding how to 'aggregate' these records are all about what assumptions you are willing to make. 

For things like the Brown Booby, we are all willing to assume that there was a single Brown Booby in the area. In fact, there may 

have been two or more, that were just never seen together. But we assume that there was one, for a whole bunch of reasons 

(often not very valid reasons, but that's a different story). 

What is important is that when you aggregate, you understand the assumptions you are making. 

To aggregate, you do some simple statistics on the set of counts for a given species within a particular date range and geographic area. 

The assumptions are contained in the decisions: 

- over how big an area should I aggregate?

- over how long a period should I aggregate?

- should I take a maximum count, or a sum (or something else, like a sum of some maxima at a lower level of aggregation)?

It doesn't really matter in the end as long as you specify the assumptions that you've made. If someone disagrees with them, 

or some additional information arises (e.g. photos of the 'single' Brown Booby at Canso turn out to be of two different birds! (Not

true as far as I know, before anyone gets too excited)). 

What I'd propose is that, off-line, you (and other editors) send us some short notes about what kind of aggregation you'd like to see, and

David and Lucas can as part of their grad course this term, do up a little script to summarize ebird data in a simple way, and provide

it back to you in a tabular format. Excel is great for tiny problems, and one could no doubt do something like this in Excel, but there

are better languages where we can accomplish the same thing very quickly. 

If other editors are interested, let us know. 

Phil.

--

To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsub...@googlegroups.com.
To post to this group, send email to ns-ebird-discussions@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/3c50d543-b18a-4c4a-aa1b-2321b019e476%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

 

--

Phil Taylor
Bird Studies Canada Chair in Ornithology
Acadia University, Wolfville, NS.

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

To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsub...@googlegroups.com.
To post to this group, send email to ns-ebird-discussions@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/CAA1OBY%2BCZGdgmNwbQ7CgOsrUiCm7%2B9HG%3DbF8dtFq0fk9JiLXvw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Dominic Cormier]

Hi All,

Before this discussion goes any further, Ian who is the chief editor for the NSBS magazine as been discussing better ways to present eBird data for the NSBS sections. Clearly this discussion is of a timely manner. We plan to try and meet in the near future with the editorial team and section editors to discuss these. I was going to suggest some additional pre-packaging of eBird data to be submitted to section editors. It seems like Phil and I are on the same page without having discussed this. I think it is possible to set up a standard script that can be used on Keith's extracted eBird data, passed to Phil (or Dave or Lucas or even myself), and then passed to the section editors. That saves SO much time for folks like John, and any other folks that don't have the best software tools to accomplish these tasks. 

Unless anyone has objections, I would like to take this discussion off here and save it for the NSBS newsletter team, and bring in Phil and co. to help with data processing if that is desired among the section editors.

Regards,

Dominic

--

To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsubscrib...@googlegroups.com.

To post to this group, send email to ns-ebird-discussions@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/3c50d543-b18a-4c4a-aa1b-2321b019e476%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

 

--

Phil Taylor
Bird Studies Canada Chair in Ornithology
Acadia University, Wolfville, NS.

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

To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsubscrib...@googlegroups.com.

To post to this group, send email to ns-ebird-discussions@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/CAA1OBY%2BCZGdgmNwbQ7CgOsrUiCm7%2B9HG%3DbF8dtFq0fk9JiLXvw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--

Phil Taylor
Bird Studies Canada Chair in Ornithology
Acadia University, Wolfville, NS.

--

You received this message because you are subscribed to the Google Groups "NS eBird discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsub...@googlegroups.com.
To post to this group, send email to ns-ebird-discussions@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/CAA1OBYLELUa0fDBatfGZtjC0nyjALVs%3D2kcHSw%3D%2BacHcquQ6HQ%40mail.gmail.com.

[Keith Lowe]

Not from me John.  Not only would have I grouped the shared checklists into one record, I don't send the observed ID. I send the observer name and code as defined by Ian but not the eBird ID

[Keith Lowe]

On Thursday, February 16, 2017 at 10:48:50 AM UTC-4, Phil Taylor wrote:

To aggregate, you do some simple statistics on the set of counts for a given species within a particular date range and geographic area. 

The assumptions are contained in the decisions: 

- over how big an area should I aggregate?

- over how long a period should I aggregate?

- should I take a maximum count, or a sum (or something else, like a sum of some maxima at a lower level of aggregation)?

 It would be good as a supplemental dataset but I doubt many section writers would want their only source of data to be aggregated to this degree. I don't write anymore because the data keeps be busy enough but I would not have been interested in this myself. If you aggregate to this level, it would not have the hotspots, the observers, the dates, the individual counts and there would be no way to determine high counts or first observations for a species, all which most writers frequently write about.

[Phil Taylor]

The point is that you can aggregate to whatever level you want, and get summaries of whatever you want at the scale of the aggregation. You could aggregate at multiple levels. Whatever any individual editor wanted. 

So, for example, you could summarize how many different people saw the booby, at how many distinct locations (e.g. different counties) on how many days and what times of day, and how many people saw it more than once, and how many took a picture and so on ... 

So I can imagine that some might indeed want it!

Cheers, Phil.

--

You received this message because you are subscribed to the Google Groups "NS eBird discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ns-ebird-discussions+unsub...@googlegroups.com.
To post to this group, send email to ns-ebird-discussions@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/ns-ebird-discussions/2502ba76-6e6a-4c25-b212-c7a1320c20f2%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.