Page 1 of 1

Size matters

Posted: 22 Dec 2010, 09:16
by Jools
On the families, family and genus pages, put an average maximum size for the species encompassed to show average size of all catfishes, catfish in the family and then genus.

Jools

Re: Size matters

Posted: 22 Dec 2010, 10:42
by MatsP
Would that be average, median or mode? Quite often it turns out to be a similar thing, but sometimes it makes a difference.

--
Mats

Re: Size matters

Posted: 22 Dec 2010, 11:07
by Jools
Mean. Median is a pain (unless a native SQL command exists - which it may well do). Mode might not exist in a set. Range might also be useful.

Jools

Re: Size matters

Posted: 22 Dec 2010, 11:29
by MatsP
It appears that median is indeed a bit more complex to calculate. There's a SQL query on http://dev.mysql.com/doc/refman/5.0/en/ ... tions.html, see the first comment, that claims to calculate median.

--
Mats

Re: Size matters

Posted: 22 Dec 2010, 11:57
by Birger Amundsen
Way over my head, to much technical English :d

Re: Size matters

Posted: 22 Dec 2010, 12:02
by MatsP
Actually, the average should be counted on vales not zero, as for many familes there are several species with no length recorded - that would skew the results.

--
Mats

Re: Size matters

Posted: 22 Dec 2010, 12:23
by Jools
Birger Amundsen wrote:Way over my head, to much technical English :d
I know, the results might be quite useful though.

Jools

Re: Size matters

Posted: 22 Dec 2010, 12:44
by Birger Amundsen
Bet it is :d

Re: Size matters

Posted: 22 Dec 2010, 13:13
by MatsP
Would it make sense to do a slightly more advanced model:
Split the sizes into groups, and draw a histogram representing their respective sizes. I think this could the same model as the current length measurement - 1..100mm, 101..1000mm, 1001 and above.

--
Mats

Re: Size matters

Posted: 01 Jan 2011, 13:14
by Jools
This has now been added to the family pages. For the Loricariidae for example, we now see: Smallest 27mm, largest 1000mm, average 176mm, most commonly 100mm. All SL.

Jools