2010-11 Shot Chart File

[DAnalytics]

I'm not sure if this data is already widely available, but I just published the 2010-11 NBA Shot Chart files to my website. I'm looking to improve the format so let me know if you have any suggestions. In the future, I plan to include shot type (ie. hook shot, jump shot, etc). I think some of the player IDs are missing as well right now.

http://www.infinitedemand.com/data/

[J.E.]

Cool stuff

I only see the same huge number in all "id" fields though. And what is your "pid"? since it doesn't seem to be bbv player id

[EvanZ]

Thanks! Echoing Jerry's comment, it could be even more useful if you use the same Player and Game ID's as basketball-value. A quick regression give -1% (FG%) for every increase in 1ft shot distance:

Code: Select all

Call:
lm(formula = made ~ distance, data = shotchart)

Residuals:
    Min      1Q  Median      3Q     Max 
-0.5845 -0.4436 -0.3228  0.4658  1.0796 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)  0.5844621  0.0017526  333.48   <2e-16 ***
distance    -0.0100623  0.0001104  -91.12   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Residual standard error: 0.4887 on 211986 degrees of freedom
Multiple R-squared: 0.03769,	Adjusted R-squared: 0.03768 
F-statistic:  8302 on 1 and 211986 DF,  p-value: < 2.2e-16 

[DAnalytics]

Right now it uses ESPN player IDs. In the future, I'd like to create a lookup table to go back and forth between the IDs on different sites. I'll take a look at what BBV uses. I'm hoping to soon get this together with PBP data in a similar format to what BBG uses.

Oh, and id is a shot ID. It is the game ID combined with a shot number. Not really important.

[Crow]

If you create the lookup table with player ID and this data is meshed with PBP basketballvalue data that would be quite useful.

In general...

and perhaps it would be useful to some degree specifically here to one or more of Evan and his EZPM, bbstats and DSMok1 and their statistical +/- models, to Rhuidean and his lambdaPM, to Jerry and his set of RAPM products or future versions of these or other models by other authors (and perhaps with versions of meta-metrics later).




And going even further (if you want to read it)...

with compiled summaries of PBP shot location data for players when they are on the court in stints of 4 or 6 minutes or longer (especially if a simple and flexible compiler interface were created), folks could simply compare that raw on the court data to Jerry's Adjusted PPS RAPM values for players and compare the shot location mixes and accuracy when a player is on the court to when he is off or league average. This could yield some thoughts about how much different shot mix distributions play into producing positive or negative player level Adjusted PPS.

(And perhaps it could also be done for specific big minute lineups if Jerry wants to produce Adjusted PPS factor level data for lineups later).

Potentially one could even sort various distributions of 3 pointers, mid-rangers and inside shots in the longer stints into a modest set of shot mix bins (maybe a simple matrix of high, average and low for each shot type) and call them "shot mix strategies" and compute a sort of Adjusted rating for these basic shot-mix strategies at various levels- player, lineup level, team (and coach) and league level. (For a season, a part of a season if a trade or major coaching / player management decision occurs or multi-season.)

Adjusted ratings for these basic shot-mix strategies would be another split beyond player and lineup roll-ups (that include various shot mix strategies) that when viewed on their own and in comparison with the other statistical and Adjusted products might help better understand or at least guess at why certain players, lineups, shot mix strategies and combinations of them were succeeding or failing on shot efficiency and overall efficiency. Sometimes shot-mix strategy (or outcome in the face of defensive strategy) may be pivotal.

To a lesser degree of accuracy one could also look at smaller sample splits against different opponent types or with other specified criteria. One might even look at the Adjusted defensive impact on a team from these different offensive shot mix strategies.

With the ability, time and interest to want to see what it would show, one could link up and compare datasets and push ever further with the analysis and maybe get greater understanding and edge on the competition.

[wilq]

I'm not sure if this data is already widely available, but I just published the 2010-11 NBA Shot Chart files to my website.

That's awesome and you have just saved me some time
Can I use this file for a couple of blog posts?

[DAnalytics]

I'm not sure if this data is already widely available, but I just published the 2010-11 NBA Shot Chart files to my website.

That's awesome and you have just saved me some time
Can I use this file for a couple of blog posts?

Sure. I just uploaded the updated 2010-11 file with the Game 6 shot chart data.

[wilq]

Can I use this file for a couple of blog posts?

Sure. I just uploaded the updated 2010-11 file with the Game 6 shot chart data.

Thanks for the fun! Even though it wasn't exactly edgy analysis...
http://weaksideawareness.wordpress.com/ ... -of-shots/
http://weaksideawareness.wordpress.com/ ... r-30-feet/

[wilq]

Can I use this file for a couple of blog posts?

Sure. I just uploaded the updated 2010-11 file with the Game 6 shot chart data.

NOTICE: This domain name expired on 07/14/2011 and is pending renewal or deletion.

Do you plan to renew it? What will happen with those files?
Will you publish them somewhere else? If not, can I do it?