01 February 2016

County rehash browns

It's February!  And that means I updated my county list map (via openheatmap.com).  Here's how the map stands currently:

It might be hard to see the differences from the older version without having a side-by-side comparison.  However, I'll fill you in.  See the trail of blue leading from Missouri to south Texas?  THAT happened this month.

So in looking at my numbers, here are some basic stats.  First, eBird shows I have lists for 714 counties (or parishes, boroughs, census areas, etc).

The sum all of my county ticks together is north of the 13k mark.  This is essentially adding up every county list I have on top of each other.

So when I calculate a simple average number of species per county, I end up with roughly 18 species.  This number is pretty low due to me having a lot of counties with 1 or 2 species on the list (thanks to drive-by county listing).

So how many counties do I have with 1 species?  Or 2?  I looked at that as well:

1 species = 108 counties
2 species = 78 counties
3 species = 88 counties
4 species = 74 counties
5 species = 42 counties

That totals 390 counties (or about 55% of my counties).

I suppose it's not too surprising that 1-bird-counties is the highest category... but it's also a bit depressing that it couldn't be 5 or something higher.

So I like to think about goals... what is the best way to summarize these data?  Total ticks?  Total counties with lists?  Averages?  Modes?

IMO, it shouldn't be "number of counties" because, as I've shown, many of those county lists only have 1 species on it.  That's not impressive!  It's especially not impressive if you aim to get at least one bird as you drive through... and then trail off and not continue to add birds.  It adds color to the map, sure, but that display is nothing more than an empty shell of tiny-tick counties.

Maybe look at total county ticks?  I like that idea but I think the spatial side of things should have some weight too.  Otherwise, you can have a handful of hella-birded counties and have almost no spatial coverage across the country.  Whatever the measure, I think area (or number of counties) needs to be included.

I know others look at, starting from the bottom of the eBird county lists, consecutive numbering, starting with counties with 1 bird, then 2, then 3, and so on up the list.  How high does yours go before there is a break in the numbers?  I had never thought to look at my data that way but I checked recently and my first break is at 37.  In other words, I don't have a county list with exactly 37 species on it.  While this is interesting, I feel like one could then try to fill gaps by adding up to a total you need... and then cutting it off (and then that would leave ANOTHER hole where that number used to be).  Anyway, it's interesting to look at once but I doubt I'll keep up with this method.

I think a worthwhile goal is to increase your mode (or median for that matter).  Of course, this is only impressive if you have a lot of counties to work with.  In other words, if you have only 10 counties listed in eBird and your mode is 8, well, big whip.  My mode is 1 (sucks).  It'd be fun if I got it to 5 though!

Another valid goal might be to see how many counties you can get to the 100 species mark and beyond.  I'm sadly lacking with this method; I have only 30 counties in the 100+ category.

I'm sure there is some equation out there that might hit the nail on the head... but I'm not sure I know it.  In fact, I'm guessing one of my friends has already done this.  Ideas???