OK, I have updated the spreadsheet on Google Docs. As the link is now buried on page 4, here it is again. I'll also edit the OP and put it there, once I finish this post.
docs.google.com
I've added a Totals page at the end (scroll the worksheets all the way to the right) so you can see at a glance the totals for each sector and star class.
So, are there any general observations to be made from the data?
Number of Terraformable Metal-Rich worlds discovered: 0 This should not have been a surprise, since only a dozen or so of the things have ever been found and reported to EDSM, making them twice as rare as GGGs. Still, I left a column in the spreadsheets for them, in case I found one.
Number of GGGs discovered: 0 This also should not surprise, though I deliberately chose Thraikoo as my first destination as it was in the "Perseus Crags" district, the approximate location of Supermoose's Lost Planet, in the hopes of rediscovering it, or at least picking up on Supermoose's trail. Still, no luck, there or elsewhere.
Rarest Planet type actually detected: Helium-Rich Gas Giants. Again, probably no surprise to anyone, but at only 21 HRGGs detected across 7000 star systems, the odds of finding one just by hoping about randomly are almost but not quite zero. Needless to say, no Helium Giants were found, either. Of course, it's well known that if you find one HRGG, you can hop all over the same boxel and you're likely to find more. But this survey was a no-boxel-hopping zone. In the number 2 rarity spot were Water Giants at 57 followed by Earth-likes at 70 and Class V Giants at 79.
Number of Notable Stellar Phenomena and Surface Bio Signals detected: 0. I wasn't really looking for these so can't rule out the possibility I overlooked them, but I don't think any of my surveyed systems or sectors fit the criteria for detecting vacuum-based lifeforms.
Observation: Water Giants are about as rare as Earth-likes, but only in AFG stars. Water giants do not seem to occur in B and hotter, or K and cooler, stars; at least zero were detected in this Project. This creates an interesting disparity at K-class, where Earth-likes are still relatively common but Water Giants nonexistent.
Now, to some of the hypotheses raised in my OP.
The WW:AW:ELW ratio. The raw stats are 751:101:70. This breaks down to 10 : 1.34 : 0.93, or 10.72 : 1.44 : 1. Ammonia worlds are clearly more common than Earth-like in the galaxy, as they consistently scored higher counts in all star classes G class and cooler - which is by far the majority of stars in the galaxy. Whether the actual galactic ratio makes it as high as 10:3:1, due to the preponderance of M-class stars, is not yet determinable (see "The Next Step", below).
The abundance of life-bearing star systems in the galaxy. The survey found 1450 out of 7000 star systems contained at least one life-bearing planet. That's one out of every 4.82 systems. My initial estimate of "about one in five star systems has at least one life-bearing planet of some kind in it" would seem to be validated, however... life was far less abundant in M-class stars than I had originally surmised. As K, M and L class stars form the vast bulk of stars in the galaxy, I strongly suspect that once stellar abundances are factored in, the probability of life will drop considerably, to somewhere around 1 in 10.
Perpetuating the paradigm that "F class stars are the best", life was most probable around F class stars, with a probability of 34.9%, or slightly over a 1 in 3 chance. A-class are almost exactly a 1 in 3 chance, while G-class stars are at 30.8%, or just under 1 in 3. K-class had a 26% hit rate (about 1 in 4), M-class stars have just a 10% chance of hosting life, while for L class stars it falls to 2.3% (less than 1 in 40).
Colonizability. Curiously, the life-stream does not precisely follow the terraformability-stream, because it is in A-class stars that one is most likely to find at least one terraformable (or Earth-like) planet, with a hit rate of 30.6%, or three out of ten. F-class comes in second, at 27.8%, with a steady decline as you get cooler. I can only surmise that A-class stars are generally younger, and younger stars are less likely to be life-bearing.
The net result seems to be, therefore, that the answer to the question of "which stars to set the filter for" depends on what you're aiming to find. If it's Maximum Earthlikes, then set filters for F. If it's max credits, then set filters for A. Setting filters for both A and F ought to optimize for both targets, without diminishing the primary target too much.
Gas giant abundances. With 6595 Sudarsky-rated gas giants detected, there's some good data to be had here. The overall breakdown is 2562:732:2562: 660:79. It's quite astonishing that Class 1 and Class 3 giants have exactly the same counts. However, as noted numerous times now, the dominance of cooler stars in the galaxy, combined with the dominance of Class 1 giants in all star types of G-class and cooler, means that Class 1 giants are clearly the most common giant type in the galaxy. And the rarity of Class II giants - in no star type are they found to the the most common - perhaps partially explains why the payout for Class II are an order of magnitude higher than any other gas giant type.
Most common bio-solvent: water. I had speculated that perhaps ammonia-based biomes might out-number water-based ones in the galaxy, if they were more common in cooler stars. Turns out, they're not. For gas giants, water-based life outnumbers ammonia-based life in all star classes. And water worlds out-poll ammonia worlds in all star classes too. No wonder the Thargoids hate us - they've got an ammonia inferiority complex.
The Next Step? A lot of these calculations are of little use without some hard data on the relative abundances of the different star types across the galaxy as a whole. There are some stats that can be pulled form EDSM, but I'm always wary of EDSM bulk data simply because we don't know how big an effect that player-filtering is having in skewing the data, towards increases counts for F-class stars and reduced counts of L-class stars, for example. So I plan on plotting a few un-filtered trans-galactic transects, get a count of several thousand stars, and see what star classes come up. The good news is, I don't even have to physically travel these transects - I can do this transect-plotting from the safety of a spacedock, and just note down what star classes the route-plotter throws at me.
