Results 1 to 20 of 36
Thread: Stats! - Charts and graphs
01-10-2013, 09:54 PM #1
Stats! - Charts and graphs
I've recently got my hands on visualisation software with which I can play around with some relatively large data sets. Therefore I rushed off and imported some of the Planetside 2 data and had a little play.
First up my weapon data from www.planetside-intel.com
As can be seen here there is a linear correlation between kills and points scored with the H-V45 having the largest score. The outliers above the trend line are mostly anti vehicle weapons with the large brown/green box illustrating a very high degree of accuracy with the Lancer and a good accumulation of points.
A second graph shows Rock Planet Shotgun outfit data from www.planetside-universe.com.
This graph shows all the VS RPS players who have logged in between 2 Sep 13 about lunchtime today. The shows a positive correlation between battle rank and score per minute. Additionally, I believe there to be a weak correlation for those with higher battle rank to have a better KDR.
Anyway, I'm willing to put together some more graphs and charts with all sorts of different data on the axes/ colours/sizes of spots/etc. and, if I learn how, add trendlines. Other charts types are available but I've got a soft spot for scatter (see above/vide supra). Use your imagination and I'll see what I can do.
Just to manage expectations I can only do this on my work computer so the throughput may be lower than people would like.
01-10-2013, 11:13 PM #2
- Join Date
- May 2013
Huh. That's really cool. Anyone else find themselves yet?
Also fun is guessing which things are other members. The Sexiest Giraffes are pretty easy to find, some of the higher BR members, but after that it can be tricky.
01-10-2013, 11:27 PM #3
- Join Date
- Jan 2013
Is it just me or are the balloon leaders tending towards higher rank compared to score/min? Makes sense.
Really like the use of different shapes, haven't seen it done much before but it works well with the ranks. Also highlights the few accomplished 'new members', who should probably be inducted and promoted, and the two anomalous lolpodders on the far right.
Last edited by Rizlar; 01-10-2013 at 11:31 PM.
02-10-2013, 11:27 AM #4
Fascinating, like staring into a fire.
I wish someone could give us a sense of how playstyle relates to that second chart. Rizlar's comment about lolpodders is definitely heading in the right direction but I think besides the obvious it could be hard to unpick. My hunch is that those above the x=y line are more team focussed and gain more xp from support roles vs the lone wolves and combat maniacs below the x=y line. But that is a hunch, may be wrong and I have no evidence to support it.
02-10-2013, 11:49 AM #5
Great work, Rizlar!
SPM is a weird metric. Besides the obvious, that it is influenced by boosts and membership, there is a historic component here: older players may generally have lower SPM than equally performing newer players, as for long vehicles gave the same XP as infantry, no ribbon XPs, no alerts, etc.
02-10-2013, 12:05 PM #6
- Join Date
- Jun 2011
There's some similar graphs over on the "kill stats" and "rank stats" pages over at our OutfitPoints page.
It's interesting to see those who played a lot but weren't very efficient (in terms of kills) over that time (like Eclectic or Eso) and those very much at the opposite end (InvaderZim, Kalezin - although of late, Kalezin seems to be a lot like IceTrance, in that he's a high-scoring member of the outfit who doesn't play with the outfit much)
02-10-2013, 12:17 PM #7
Well the size of the spots indicates how many kills have been made by the player so that should help although it's not the easiest thing to see. However, as I understand it (I'm prepared to be told otherwise), the score per minute is uncorrected for boosts/XP bonuses etcetera which further complicates things.
I could try to dig out some data on specific dots/players if you're interested which may indicate playstyle by correlating with weapons. I would post it with names redacted to protect the innocent unless the specific parties are happy to be named. There has been some discussion recently in the WAS thread that I'd like to visualise. I am reliant on using existing resources as named above as I have no idea how to muck around with the API to get the data out of it and I'm unwilling to learn just now.
02-10-2013, 05:40 PM #8
02-10-2013, 05:43 PM #9
- Join Date
- Feb 2013
- Stockholm, Sweden
I won't mind if you pull my data as well. I only use a few different weapons, but it would be pretty interesting to see how my support stuff (medicing, repairing, etc.) or flying stuff affects it all.
02-10-2013, 06:27 PM #10
You can extract my crappy stats as well it's not like they are hidden or that people don't know already that if you look at it from a pure stats point of view, I'm one of the crappiest players in the outfit
02-10-2013, 08:48 PM #11
02-10-2013, 10:45 PM #12
I'm a BL at battle rank 66. That could be me on the left. I know I have a low kd relative to spm. That's a result of lots of reviving, tank killing and repairing. Plus I'm a dreadful shot! Happy to have my stats analysed further, for science.
02-10-2013, 11:24 PM #13
03-10-2013, 12:00 AM #14
03-10-2013, 01:11 AM #15
I havent played in a while, otherwise there would be a random square near the bottom in the 20s 30s :P.
03-10-2013, 10:15 AM #16
- Join Date
- Feb 2012
I'm kind of in the middle there somewhere, as floating in the air for an entire evening doesn't get you much in the way of actual score. Despite playing since the beta, I only have 3.6k kills
The stat I'm most smug about is the ratio of Damage Given / Damage Taken, where I'm at about 8/1 due to a habit of killing ALL THE SUNDYS.
PPA sniping FTW.
03-10-2013, 11:41 AM #17
- Join Date
- Jan 2013
Add me to the 'happy to use data' list, though I doubt mine is particularly interesting. Or maybe it would be interesting, in that it is surprisingly average?
03-10-2013, 11:47 AM #18
This is awesome. Science! Feel free to use my data.
03-10-2013, 01:25 PM #19
Started working with Eso's data as he replied first. Things I've learned - Frag grenades screw up your data. Also people having a favourite weapon skews the data too. Hence I'm going to be working with log kills and I've deleted the data to do with frag grenades. Therefore on the scale below:
0 = 1 kill
1 = 10 kills
2 = 100 kills
3 = 1000 kills
Some interesting data around headshots. I assume it only counts headshots which kill.
Pistols are headshot beasts. The ursa being up there is interesting too.
(I don't know how to delete attachments... Pauly you're a genius)
03-10-2013, 02:12 PM #20
Frag nades and knifes only count as equipped during the animation, so a few seconds per use. Normal guns count as equipped if you're just holding them, so they have a lot more time.