The Advanced Stats Thread Episode VI: RIP To Our Databases

[Machinehead]

Leafs: *hire analytics guy*

Also Leafs: *hire bad defenseman because he was a veteran and stuff*

 

[Mac n Gs]

Here you go, nerds:

 

[Greg02]

If you're starting as a beginner, there's a bunch of reasons I'd recommend going with Python or even R over SQL. Using SQL might make sense when a database already exists and there's already a complex data model, but in the case of CSVs, there's going to be less overhead and a lot more flexibility with something like Python.

 

[silverfish]

NFL DFS scraper and charts are built for the year. Do I switch from R to Python and learn to code the right way, or do I stick with R, and try to fix my warts and actually learn what this thing does...

 

[PlamsUnlimited]

Switch so you can help me

 

[silverfish]

Switch so you can help me

Trying to keep my eye on job openings. Not really looking at the moment, but feel like the recent NHL jobs I've seen, not that there have been many, have been more Python than R.

 

[PlamsUnlimited]

Trying to keep my eye on job openings. Not really looking at the moment, but feel like the recent NHL jobs I've seen, not that there have been many, have been more Python than R.

When I read some forums, maybe it's selective memory but people like Python too. I'm sticking with it, but struggling right now.

 

[Irishguy42]

From what I've seen on twitter and around the internet in general, it does seem like there is some momentum to switching over to Python.

 

[Greg02]

When I read some forums, maybe it's selective memory but people like Python too. I'm sticking with it, but struggling right now.

What are you struggling with?

 

[Kane One]

If you're starting as a beginner, there's a bunch of reasons I'd recommend going with Python or even R over SQL. Using SQL might make sense when a database already exists and there's already a complex data model, but in the case of CSVs, there's going to be less overhead and a lot more flexibility with something like Python.

What's the context of this? I don't see how anyone could possibly recommend sticking with CSVs over a database.

 

[PlamsUnlimited]

What are you struggling with?

Trying to install a package for python. For some odd reason that I can't seem to figure out, it's giving me one of two errors and I've never had them before on it

 

[oooooooooohCanada]

Leafs: *hire analytics guy*

Also Leafs: *hire bad defenseman because he was a veteran and stuff*

Robidas was a very good defenseman for many years. Career top-4 dman, top pairing at his peak.

 

[silverfish]

What's the context of this? I don't see how anyone could possibly recommend sticking with CSVs over a database.

Less of a learning curve?

We do a lot of pretty powerful **** with Excel at work these days because my agency won't move to R. Not worth it for them to restructure everything. Boils my blood but what can ya do. My Excel game is top-notch

Robidas was a very good defenseman for many years. Career top-4 dman, top pairing at his peak.

I think he meant Hainsey

 

[Kane One]

Trying to install a package for python. For some odd reason that I can't seem to figure out, it's giving me one of two errors and I've never had them before on it

Are you using pip? What's the error message?

 

[NickyFotiu]

I don't mind having them work with prospects and player development. I find their involvement with hockey-related decisions concerning though.

Why?

 

[PlamsUnlimited]

Are you using pip? What's the error message?

Yeah it comes with the newest version, at least that's what forums tell me. I'm still quite new to stuff like this though. But it either says no such package is installed or when I tried to reinstall it using the directions the creator posted it keeps giving me syntax error. I'm not sure why if I'm doing it word for word

 

[Kane One]

Yeah it comes with the newest version, at least that's what forums tell me. I'm still quite new to stuff like this though. But it either says no such package is installed or when I tried to reinstall it using the directions the creator posted it keeps giving me syntax error. I'm not sure why if I'm doing it word for word

Check to make sure you're using the correct version of Python. The package could be Python 3 and you may be using 2.7.

What's the exact error you're getting now?

 

[Greg02]

What's the context of this? I don't see how anyone could possibly recommend sticking with CSVs over a database.

If you have to manually download CSVs in order to upload them into a DB, at this data scale you're getting no real benefit and just an added step of something to do. It's not like adding indexes is going to affect your life in any meaningful way when you're talking about KBs. You also have to think a lot more formally about your data model up front, and maintaining/updating the DB without any sort of automated pipeline is going to be more annoying for someone just starting out.

 

[Kane One]

If you have to manually download CSVs in order to upload them into a DB, at this data scale you're getting no real benefit and just an added step of something to do. It's not like adding indexes is going to affect your life in any meaningful way when you're talking about KBs. You also have to think a lot more formally about your data model up front, and maintaining/updating the DB without any sort of automated pipeline is going to be more annoying for someone just starting out.

So let's say I want to go on this website/app that uses CSV files as its database and I want to view a range of data that spans over multiple files, is the code going to iterate over these files to select the data? Or does the app iterate over all the files when the app is initially loaded?

Even if the data isn't complex at all, they should still learn SQL which is obviously an incredibly valuable skill to have.

 

[Greg02]

So let's say I want to go on this website/app that uses CSV files as its database and I want to view a range of data that spans over multiple files, is the code going to iterate over these files to select the data? Or does the app iterate over all the files when the app is initially loaded?

Even if the data isn't complex at all, they should still learn SQL which is obviously an incredibly valuable skill to have.

Sure, iterate and load into memory. It'll almost certainly take under five seconds at this data scale. What's the alternative? Download all the CSVs manually and load them into the DB individually by hand?

Obviously, the better thing to do is write a script that downloads all the data for you and systematically loads it into a database. But then we're talking about working in something like Python anyway, so might as well start there, where you can do everything SQL can do and more, and have less startup steps.

 

[Raspewtin]

Why?

Instead of bringing in fresh minds with business and management experience, constantly hiring former players that "paid their dues" contributes to the garbage OBC that hockey is today.

Cool, I doubt they have much input in decisions, but I am really tired of every position going to former players.

 

[silverfish]

Instead of bringing in fresh minds with business and management experience, constantly hiring former players that "paid their dues" contributes to the garbage OBC that hockey is today.

Cool, I doubt they have much input in decisions, but I am really tired of every position going to former players.

I'm so tired of the Rangers being mediocre, why is this always the case!

*Rangers hire former players*

Wow! So great having my favorites like Leetch back in the organization.

Rinse and repeat.

[of course this is an oversimplification, but it shouldn't be ignored.]

 

[NickyFotiu]

I think I'm in the wrong thread. I definitely want guys that can evaluate hockey players actual skills on the ice. I prefer guys with some experience in playing, scouting and or coaching the game even if that is below the NHL level. That is not to say that we can not have other outside points of view included in our management team but I would never rule somebody out because they have actual hockey experience. I would not want Leetch to come do my income taxes nor do I do want the best buy geek squad deciding which players have the best hockey skills to succeed in the NHL.

 

[silverfish]

I think I'm in the wrong thread. I definitely want guys that can evaluate hockey players actual skills on the ice. I prefer guys with some experience in playing, scouting and or coaching the game even if that is below the NHL level. That is not to say that we can not have other outside points of view included in our management team but I would never rule somebody out because they have actual hockey experience. I would not want Leetch to come do my income taxes nor do I do want the best buy geek squad deciding which players have the best hockey skills to succeed in the NHL.

Yes, as usual, the extremes are brought into the picture while completely ignoring the middle.

You're not in the wrong thread, you're just not understanding the argument.

The argument is not: "Never hire former players".

The argument is: "Let's hire people with actual, tangible, experience".

Leetch wants to work in an NHL FO? Get hired in the AHL. Get hired in the OHL. Prove you can do it.

The argument is: "Former players shouldn't be hired to NHL FOs just because they are former players".

 

[NickyFotiu]

Yes, as usual, the extremes are brought into the picture while completely ignoring the middle.

You're not in the wrong thread, you're just not understanding the argument.

The argument is not: "Never hire former players".

The argument is: "Let's hire people with actual, tangible, experience".

Leetch wants to work in an NHL FO? Get hired in the AHL. Get hired in the OHL. Prove you can do it.

The argument is: "Former players shouldn't be hired to NHL FOs just because they are former players".

Leetch's role is not defined at this point. I'd prefer he work with the players because he knows exactly what it takes to be a Cup Winning NHL Player. There is so much more than the average fan sees on TV. I have to ice my feet in a bucket for 30 minutes before my own games to have good legs. The travel alone would kill me in the NHL. I do not know how some players can play on or near some travel days. Brad Richards only won a cup 3 seasons ago so he may have more insight in to the guys playing today than Leetch. I'd be happy for him to give his opinions to Slats/Gorton/Drury/AV if he feels he has something important to share with them about a specific guy.

The guys that are given roles that are more than ceremonial usually earn them along the line at some point. Most NHL franchises have had 1000s of players so they can't just give many roles away frivolously. I'm sure the staff see something in Brad and Brian they hope can help us win in some way. If that is not the case their role will become more like Adam Graves/Rod Gilbert than Drury or just become a temporary position.