ED Astrometrics: Maps and Visualizations

I've got a question / suggestion. Do you know how to identify captured planet candidates? They're supposed to exist, and I've found a few possible ones over the years. You may have already made one, or perhaps like me you have had the idea but are uncertain about the criteria.

An example of one I'm fairly certain is a capture.

The crazy orbit isn't actually what made me notice this one, but rather that it was an anomaly within the boxel. (Out of 192 systems, there were no other Ammonia Worlds, and only one Water World (Oephail LG-Y e84 11), which also has a somewhat odd orbit and unique characteristics compared with the rest of the boxel)
 
Last edited:
Yeah, that's a really tough call. There's no "flag" to specifically mark them as such in the journals. So we would have to make assumptions based on orbital parameters, etc. Since that one has a high eccentricity, that does make it a good candidate. However those can also occur due to interactions with another body, which may have been ejected.

But looking at the composition could certainly be valid too. We just have no way to verify any meaning behind it.

So really the best I can do is make spreadsheet reports based on certain features, such as the inclination or eccentricity, and leave it to the viewer to decide what the meaning of those values are.
 
Yeah, that's a really tough call. There's no "flag" to specifically mark them as such in the journals. So we would have to make assumptions based on orbital parameters, etc. Since that one has a high eccentricity, that does make it a good candidate. However those can also occur due to interactions with another body, which may have been ejected.

But looking at the composition could certainly be valid too. We just have no way to verify any meaning behind it.

So really the best I can do is make spreadsheet reports based on certain features, such as the inclination or eccentricity, and leave it to the viewer to decide what the meaning of those values are.
Taking a stab at the composition, there are a few things we might be able to ascertain, assuming we have enough data from the boxel as a whole, and assuming we can draw some tentative conclusions from "similar" boxels. (btw, I'm operating under the assumption that you are familiar with my usage of 'boxel')

I've had a focus on helium, specifically boxels with over 29.3% (the point at which Helium Rich Gas Giants start to appear), and those below 24%. These examples are in the latter category. A common characteristic of low helium boxels is that they have relatively few bodies (I haven't found one with more than 23, out of the 800 or so I've scanned) and virtually no boxels have terrestrials aside from rocky, rocky-ice, hmc, icy, and metal-rich (with hmc being the overwhelmingly dominant type). Oephail LG-Y e has the distinction of being one of only two boxels I've so far found that has a water world, and it's the only one thus far with an ammonia. All three of these worlds have unusual orbits.

Gas giants are also very uncommon. Obviously each of these boxels has at least one, since I couldn't otherwise get any sort of helium percentage out of them. Oephail LG-Y once again has the most gas giants, with a total of 7 (out of 192 systems). So far, the higher mass code (D,E) boxels have had the lowest number of gas giants relative to the number of systems. But most of the known low percentage helium locations tend to be C's, and most of those are within a few thousand ly of The Bubble, so it's hard to draw any conclusions from that.

The point being, I'm not sure that these types of planets can form in low helium systems.

This might be outside of your area of interest, so I'll stop now. If you're interested I can provide you with my data (though it's all on EDSM, of course).
 
Last edited:
That is pretty cool. I'm always amazed at what we can derive from the data, even with such a limited view of what the StellarForge has done internally. I guess my question would be this, is there a spreadsheet or set of sheets that would be useful for this sort of thing? Per-boxel statistics or something? And if so, what statistics would be useful? I'm assuming boxels in this case refer to specific sets of "XX-X m" patterns (letter codes plus mass code), since I think that's what we all tend to refer to with the word "boxel".
 
I think per-Boxel statistics would be incredibly useful. Helium (from gas giants) has so far been the one we've focused on the most, since it's quite easy to track. But others I'm finding relevant include bodycounts (again per system. Boxels would probably benefit from knowing highest and lowest numbers), mass for various object types (this is harder, since two different planets of the same type can be quite different. I'm thinking about how to make subclasses, and how to programatically identify them), which object types exist, age (working on a hypothesis on this, but there's some evidence that age matters with regard to ring formation) and almost certainly other things. I've written my own journal parser that creates separate outputs for sectors, boxels, systems, and objects, but I'm still struggling with how to categorize large parts of the data.

Out of interest, what do you use to parse the EDSM logs? I've written a small parser, but I'm not terribly good at it.
 
I do all of the parsing and scripting in Perl. I have a script that parses the JSONs from EDSM and shoves the data into a MySQL database, and from there I can script against it to extract whatever I need. What gets tricky is that the data set is quite large, so different scripts often require different strategies for limiting the search or chunking the data. But I guess that's half the fun. :)

EDIT: For Helium, how are you determining the boxel content? Average percentage of the gas giants, or something along those lines?
 
EDIT: For Helium, how are you determining the boxel content? Average percentage of the gas giants, or something along those lines?
I'll get the helium percentage of one gas giant per system that has one and get the average per boxel by taking the helium levels of each of those systems and dividing by the number of those systems. I also record the boxel highest and lowest. I've also set my script to prefer gas giants that don't have life, whenever possible. (They will always have slightly less helium, which skews the numbers very slightly. I'm fairly sure that every system has a set helium/hydrogen percentage, and that gg w/life subtracts from that.)
 
o_O … that's interesting.
And I have several questions:

What is a boxel?
You write:
'm assuming boxels in this case refer to specific sets of "XX-X m" patterns (letter codes plus mass code)
But I hear that there is something more behind the "XX-X" for the first time. Albeit I have suspected it since I had some long term plans to analyse it.
So what exactly is that?
Also, is my assumption correct that this "works" just for names like "NGC 6067 Sector ZU-Y d31" but not "V1292 Scorpii" or "CD-23 13397"?

Secondly, last week I started an analysis of the composition of the galaxy … To realize that (even after the FSS) a lot of the data is simoly not available because it isn't scanned. And I admit that I'm one of those that play the FSS minigame rather seldom, and often not completely.
How will that affect the results?

I've heard that some commanders have scanned entire subsectors. I assume this is not the same like boxels, or is it?
Are these results "skewed" towards one "element distribution" or another?
Is the expected element distribution the same everywhere or are there local "hotspots". The latter NOT in a a system but galaxy wide at least on the scale of nebula?

More questions probably later.
 
What is a boxel?
The Sector Naming article on the DISC wiki is the best explanation.
Boxels aren't just "XX-X" as you wrote, but "XX-X n" instead. We call the n there the mass code.

Also, is my assumption correct that this "works" just for names like "NGC 6067 Sector ZU-Y d31" but not "V1292 Scorpii" or "CD-23 13397"?
Your latter two examples are imported catalogue systems, not procedurally generated, so yeah. Your first example might be off, as the "sector" named systems are all overrides of the procedurally generated names, in a certain distance of a point. So while procedurally generated sectors (for example, Synuefe, Wregoe) are boxes, override sectors (NGC 6067 Sector) are spheres. (There are times when some of them overlap though, and they take parts out of another.) They still have their original names in the game, but most of the time, that'll turn out different than the new one, so you can't assume that Happy Hippo Sector ZU-F c4-20 would be Blu Aec ZU-F c4-20.

As for the contents, it has been shown that in regards to helium percentage, and thus most likely metallicity as well, individual boxels share the same levels. AFAIK some suspicion too that systems in the same boxels share further similarities, but not definite proof. Others would be able to tell you more on this subject though.


Oh, and your remark about using the FSS seldom and not to completion: don't feel bad about that. Not only do the majority of the players do the same, the developers also said that's the usage they planned it for. Jump into system, look at the barcode to see if there are any body types you are interested in, jump out if there aren't.
 
I may play with scripting something up to look at the boxels, but I suspect I'll have to have a cutoff based on mass code or something. That is, analyze D-mass and above, or something similar. I'm just not sure yet. I want to avoid making lots of spreadsheets that have tens of millions of rows, but I'm not sure how the existing data will break down yet. Clearly there's an upper limit based on the number of star systems. Worst case scenario: if every single system were in different unique boxels, then there would be 42 million boxels in the current data. Clearly that won't be the case. But in the less explored regions it's probably common to have boxels with only one visited system. I'll probably have to play around with it to see where it might be useful to limit the data.

This will be a slow-running script, so depending on how it works out, I might have to have it do weekly updates instead of running every two days. But I'm just speculating on the outcome at the moment.
 
I may play with scripting something up to look at the boxels, but I suspect I'll have to have a cutoff based on mass code or something. That is, analyze D-mass and above, or something similar. I'm just not sure yet. I want to avoid making lots of spreadsheets that have tens of millions of rows, but I'm not sure how the existing data will break down yet. Clearly there's an upper limit based on the number of star systems.
Neat! If you want to try a test case, might I suggest the sector "Scaulua"? That's the cube the showed up left of Colonia during DW2, and I made a large scale survey of every E, F, G, and H boxel in the sector. So you'll have a bit over 4000 complete system scans to work with. And though I didn't scan 100% of every boxel (that would be... hard.) a part of my surveying pattern is to identify and scan the last numbered system in a given boxel. So you will also get an accurate system count for E and up in Scaulua.
 
Some initial tests, using Oephail and Scaulua:

Oephail has 3002 systems in 2062 boxels. Limiting to d-mass or higher returns 548 boxels. Limiting instead to boxels of 10+ systems returns 11 boxels.

Scaulua has 5999 systems in 2335 boxels. Limiting to d-mass or higher returns 1684 boxels. Limiting instead to boxels of 10+ systems returns 67 boxels.


I may have to just use the system count as a threshold for it to have interesting results, unless we want to tighten the mass codes to e+ or something like that.

Here's the CSV output I have so far for just "Oephail LG-Y e":

Code:
"Boxel","Systems","Age Avg","Age Min","Age Max","Hydrogen Avg","Hydrogen Min","Hydrogen Max","Helium Avg","Helium Min","Helium Max","Total Stars","Avg Stars","Min Stars","Max Stars","Total Planets","Avg Planets","Min Planets","Max Planets","Total Gas Giants","Avg Gas Giants","Min Gas Giants","Max Gas Giants","Total Terrestrial Bodies","Avg Terrestrial Bodies","Min Terrestrial Bodies","Max Terrestrial Bodies","Avg X","Avg Y","Avg Z","Min X","Min Y","Min Z","Max X","Max Y","Max Z"

"Oephail LG-Y e","192","206.84","2","1726","76.09","76.0689","76.13","23.89","23.86","23.93","334","1.74","1","7","36","0.19","0","5","7","0.04","0","1","29","0.15","0","5","7534.04","51.78","17092.64","7455.22","-23.44","17016.00","7612.81","133.81","17174.30"
EDIT: Added stellar ages, and min/max hydrogen and helium.
 
Last edited:
Some initial tests, using Oephail and Scaulua:

Oephail has 3002 systems in 2062 boxels. Limiting to d-mass or higher returns 548 boxels. Limiting instead to boxels of 10+ systems returns 11 boxels.

Scaulua has 5999 systems in 2335 boxels. Limiting to d-mass or higher returns 1684 boxels. Limiting instead to boxels of 10+ systems returns 67 boxels.


I may have to just use the system count as a threshold for it to have interesting results, unless we want to tighten the mass codes to e+ or something like that.
E & up might be interesting. I can tell you that I've 100% scanned 5208 systems in Scaulua. 5209 if you count Scaulua OY-R e4-15, but even though I went back there a second time to be sure, it never outputted the FSSAllBodiesFound in the journal no matter what I did.

You might try a somewhat lower number of systems per boxel in Scaulua. My goal was to have at least three systems with gas giants per boxel, since I did that on my own and I'm happy not to still be there. :D There were only 2-5 systems per boxel on the bottom layer, and a few of those didn't have gas giants at all. I believe my minimum goal was 5 systems where possible, though nearly half of the E boxels in Scaulua had at least 7.
 
OK, here's a test run across the whole galaxy, limited only by mass code, e+. The mass code limitation allows the initial lookup of system IDs to be filtered by name, making the script run in a pretty short time. This results in a spreadsheet of 376,367 boxels:

Looking at it now.. something odd is happening with the helium / hydrogen average, at least in some of the boxels... How is the average being calculated? An example of a problem boxel is Wepooe EA-A e, where the average is lower than both the maximum and the minimum.
 
A request: Would it be difficult to add a column for the system with largest number of bodies in a given boxel? (Not the name of the system. Just the number of bodies in the one that has the most, both stars and planets) I have a hypothesis I want to test.

Another useful field would be one that just shows the boxel letter. (To filter out the H's, E's, whatever)
 
Last edited:
Looking at it now.. something odd is happening with the helium / hydrogen average, at least in some of the boxels... How is the average being calculated? An example of a problem boxel is Wepooe EA-A e, where the average is lower than both the maximum and the minimum.
That's strange. Probably a stupid typo in the code, or something, or referencing the wrong variable. I'll look.

A request: Would it be difficult to add a column for the system with largest number of bodies in a given boxel? (Not the name of the system. Just the number of bodies in the one that has the most, both stars and planets) I have a hypothesis I want to test.

Another useful field would be one that just shows the boxel letter. (To filter out the H's, E's, whatever)
Yep, I could add total bodies. I'll probably add all three, average, min, max.

For boxel letter, you mean the "XX-X" or the mass code?
 
Top Bottom