![]() |
Coming This Summer: The SHL as Shown in R - Printable Version +- Simulation Hockey League (https://simulationhockey.com) +-- Forum: League Media (https://simulationhockey.com/forumdisplay.php?fid=610) +--- Forum: SHL Media (https://simulationhockey.com/forumdisplay.php?fid=46) +---- Forum: Graded Articles (https://simulationhockey.com/forumdisplay.php?fid=545) +---- Thread: Coming This Summer: The SHL as Shown in R (/showthread.php?tid=94781) Pages:
1
2
|
Coming This Summer: The SHL as Shown in R - GCool - 03-30-2019 Hey guys, As a lot of you know, I'm currently studying biostatistics - which is a fancy word for applied statistics. I don't know much bio. For now, I've been learning a lot of theory as well as a new programming language called R (yep, just the letter). R is an open-source, free-to-use programming language that's geared towards statistical analysis. For programmers out there, it's sort of a hybrid of SQL and Python with a lot of unique functionality contributed by statisticians and coders from all over the world. The key advantage in working in R, at least from my industry's perspective, is that graphs are a breeze - and they're beautiful. I'm sure some people can/will attest that this beats the shit out of what Excel can do. Right now I'm learning how to create good data models in one class while learning how to visualize them in this software in another. It's been very rewarding in spite of it being a lot of work since it's a 2 hour commute to school plus working full time an hour from home and then managing to make my wife really happy. It's hard out there for a pimp like me. Anyway, I wanted to show you guys what R can do, so I took the Draft Class / TPE list from Luke and the Budget sheet and made this really quick: ![]() I shit you not, this took about 15 minutes. Granted, I've been learning this in school so it's fresh in my head, but it honestly wasn't bad. We can gather a few things from this graph -- mostly that Esa is the reason everyone else's graphs look so skewed. I could independently shift everyone's X axis, but ... eh. Then it'd take about 25 minutes. We can also see at what stage multiple teams are in development - I mean, look at Chicago. They got one lonely guy at 1500 TPE, and the rest are very far right on the X axis (read: young) and low in TPE right now. I wanted to create this preview to ask the audience what else they'd want to see visualized, and I want to see if I can actually handle it. There are a lot of great color pallets, different graphing tools, and of course ... regression models. One model I really want to explore is the value of the "OVR" rating versus certain ratings. I know there's been some exploration into this, after talking to old veterans for as long as I have. The consensus is that it doesn't matter, so obviously nobody will care if I tear that axiom apart and explore what OVR is actually worth. But uh, yeah, that's it. Excited to hear some questions. Maybe I can turn this into a presser with the right questions. RE: Coming This Summer: The SHL as Shown in R - luke - 03-30-2019 Manhattan young and high TPE RE: Coming This Summer: The SHL as Shown in R - WannabeFinn - 03-30-2019 03-30-2019, 01:26 PMluketd Wrote: Manhattan young and high TPEYour graph also tops out at 1,000 TPE lol RE: Coming This Summer: The SHL as Shown in R - luke - 03-30-2019 03-30-2019, 02:08 PMWannabeFinn Wrote:03-30-2019, 01:26 PMluketd Wrote: Manhattan young and high TPEYour graph also tops out at 1,000 TPE lol 1100 bud. Can only go up with everyone approaching and passing 1000+ RE: Coming This Summer: The SHL as Shown in R - .bojo - 03-30-2019 Finally, good to see somebody doing some R work for shl. I always thought it'd be perfect to practice on. Still want to get into it myself. RE: Coming This Summer: The SHL as Shown in R - GCool - 03-30-2019 03-30-2019, 02:50 PM.bojo Wrote: Finally, good to see somebody doing some R work for shl. I always thought it'd be perfect to practice on. Still want to get into it myself. I assume it would be pretty easy for you to pick up, honestly. And yeah since we learned how to take in CSVs, I've been thinking up ideas for sim data. Taking any and all suggestions though! RE: Coming This Summer: The SHL as Shown in R - Tomasnz - 03-30-2019 Okay I love the work you have done. And I get this is a first of many steps. But for the love of God please normalise those y-axis O,o ![]() Great work so far. RE: Coming This Summer: The SHL as Shown in R - GCool - 03-30-2019 03-30-2019, 02:55 PMTomasnz Wrote: Okay I love the work you have done. I laughed so hard just now. I knew that'd throw some people off. But yes, when I'm actually running this, both axes will be normalized. I promise. It' s actually an additional step to do it the way I did; by default they have the same axis (like the x is right now, thanks old man Esa). But this way everyone feels a little better about their team at a cursory glance. ![]() RE: Coming This Summer: The SHL as Shown in R - Tomasnz - 03-30-2019 03-30-2019, 02:58 PMGCool Wrote:03-30-2019, 02:55 PMTomasnz Wrote: Okay I love the work you have done. Nooooooo! Your turning the stats into lies! But fair. As long as you know your breaking good presentation rules I guess that is sort of okay . Still would love to see what normalised graphs look like . Hint hint ![]() RE: Coming This Summer: The SHL as Shown in R - GCool - 03-30-2019 03-30-2019, 03:05 PMTomasnz Wrote:03-30-2019, 02:58 PMGCool Wrote: I laughed so hard just now. I knew that'd throw some people off. It'd probably look like this: ![]() RE: Coming This Summer: The SHL as Shown in R - BDonini - 03-30-2019 This is awesome. I've been looking to learn R on the side but haven't had time during the semester...hopefully once summer hits I can start getting into it. Not sure if you know Sean Tierney but it would be cool to see Sean Tierney type visuals for the advanced stats that @aaronwilson and @Baelor Swift created. RE: Coming This Summer: The SHL as Shown in R - GCool - 03-30-2019 03-30-2019, 03:47 PMBDonini Wrote: This is awesome. I've been looking to learn R on the side but haven't had time during the semester...hopefully once summer hits I can start getting into it. Not sure if you know Sean Tierney but it would be cool to see Sean Tierney type visuals for the advanced stats that @aaronwilson and @Baelor Swift created. That's one of the goals for sure! I just recently was told to look into his stuff - I will give it a look soon ![]() RE: Coming This Summer: The SHL as Shown in R - Tomasnz - 03-30-2019 03-30-2019, 03:40 PMGCool Wrote:03-30-2019, 03:05 PMTomasnz Wrote: Nooooooo! Your turning the stats into lies! ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() RE: Coming This Summer: The SHL as Shown in R - awils13 - 03-30-2019 Awesome work, love seeing stuff like this RE: Coming This Summer: The SHL as Shown in R - Zoone16 - 03-31-2019 I like this a lot. Shows a lot and tells you what team is at what stage of their development. Next Step: Take the normalized graph and color thedots by position (F,D,G) |