04. The discrepancy between Observations on obs search:

i'll continue being a broken record on the other major discrepancy/inconsistency that causes differences between different lists and that is the illogical, arbitrary rule that cultivated/domestic plants/animals shall be declared as unable to be research grade. these records contain important ecological information and are often the entrée to iNaturalist for many urbanites who want to know what 'that plant' is. there is no requirement from GBIF that they only accept 'wild' organisms, only that they are identified as wild or not. but this becomes another source of discrepancy and confusion in bioblitzes and people looking at places versus place-bounded projects or 'species guide' versus 'observations' without realising one needs to drill down deeper in order to figure out the reason for differences (which no one will know to do). My colleagues at NatureWatch NZ agree. And as Charlie or Wolfgang said, these numbers need to be consistent and all visible in all relevant pages; that would solve the problem, so that users know what the different numbers mean - which are all legitimate ways of presenting the data. cheers c

On Thursday, May 25, 2017 at 6:43:25 AM UTC+12, Scott Loarie wrote:
Hi Mike,

The discrepancy between Observations on
obs search:
the people leaderboard:

is that obs search shows 'verifiable' (ie needs_id+research) by default where as the people leaderboard shows all obs (ie casual+needs_id+research)

You can see that if you uncheck 'verifiable' on obs search, they match:

The discrepancy between Species on obs search and the people leaderboard is due to the 'verifiable' discrepancy described above
but also the way 'species' are counted.

Here: http://www.inaturalist.org/observations?d1=2017-04-01&d2=2017-04-30&place_id=any&verifiable=any&view=observers
they are counted as 'leaves with ssp rolled up into species'
which means all the tips of the tree where ssp are counted as their parent species

Whereas, here: http://www.inaturalist.org/people/leaderboard/2017/4
they are counted as 'all taxa of rank species or finer'

For example, obs of the following
Taricha torosa
Taricha torosa
Taricha torosa torosa

Would be counted as 2 'species' on both obs search and the people leaderboard, but for very different reasons
Obs search: Taricha torosa (3 obs) + Ensatina (1 obs)
People leaderboard: Taricha torosa (2 obs) + Taricha torosa torosa (1 obs)

I'd be in favor of making the people leaderboard page reflect obs search so this is consistent, but curious what others think


On Tue, May 23, 2017 at 11:13 AM, Mike Burrell @ontario.ca> wrote:
Hi there,

If one goes to the observations page and clicks the observers button and then filters it for the current month you see a list of the top contributors that you can sort by either Observations or Species (http://www.inaturalist.org/observations?d1=2017-05-01&d2=2017-05-23&place_id=any&view=observers). Currently the top five contributors for observations are:
1 ck2az 3,310
2 erikamitchell 2,245
3 srall 1,806
4 momoto-erick 1,501
5 anudibranchmom 812

One can also access this same information by clicking the people tab and then scrolling down to the leadboards (http://www.inaturalist.org/people). There the top five contributors are:

  1. ck2az 3310 (same)
  2. erikamitchell 2262 (7 more)
  3. srall 2042 (236 more)
  4. momoto-erick 1501 (same)
  5. belyykit 935 (not in top 5 on observations page)

For species they are:
Observations page:
1 srall 426
2 erikamitchell 399
3 sambiology 378
4 catchang 296
5 mako252 279

People page:
1 srall 524 (98 higher)
2 erikamitchell 400 (1 higher)
3 sambiology 369 (9 lower)
4 catchang 292 (4 lower)
5 bouteloua 288 not in top 5 on observations page

Why is there a discrepancy between these two pages that should theoretically be the same. There doesn't seem to be consistent with one page having higher numbers than the other (as I would expect if it were just one page being updated more recently than the other). Are the two pages actually calculating two slightly different figures? Does one of them only count research grade or verifiable observations?



Publicado el 04 de septiembre de 2018 a las 08:53 PM por ahospers ahospers


Orcon, WTW, 90%+, paar duizend aan spullen, paar duizend aan installatie (afhankelijk van wat je zelf kan), ruime eengezinswoning woning, jaren 1960 - volledige renovatie naar niveau 2025


Anotado por ahospers hace mas de 3 años

Añade un comentario

Entra o Regístrate para añadir comentarios