|Title:||I analyzed James Harden's performance in every NBA city to see if I could find a correlation between his box score and the city's average strip club rating.||Side Notes:|
|Intro:||Everyone knows James Harden has a particular affinity for the Canadian ballet, aka Strip Clubs. After the Rocket's dismal performance in Miami last week, and the city's reputation for high quality tit-shacks, I became increasingly curious to see just how much James Harden's vice affects his game. So here we are, I spent the better part of the work week on this, hope you enjoy! |
Hypothesis: James Harden's box score declines in cities with high quality strip clubs
Test: Analyze James Harden's performance in every NBA city and correlate with those cities' reputation for stip clubs to see if there is any discernable relationship.
|did not distinguish between gay/straight strip clubs|
did not distinguish between cabarets, strip clubs, men's clubs
did not disqualify paid listings on google. If a paid listing was the top result, it was included in the average
when searching brooklyn, some strip clubs came up that are in Manhattan, did not distinguish between these. Just took the top 10 that show up when searching "[city] strip clubs"
I did not factor in Team performance (win vs loss) into Harden's "poor performance analysis"
considered adding an "Opponent Strength" coefficient to the performance analysis based on either end of season Defensive Ranking or on end of season W/L totals. E.g., if Phoenix had the 29th ranked defense, Harden's performance would be expected to exceed the season average, so if he only met the season average it would be considered a "sub-par" performance. I decided against this because it leaves too much room for me to massage the data. I'd have to spend hours developing a consistent coefficient and even then it'd be subject to scruitiny.
|Methodology/ Steps:||First I extracted all of James Harden's game logs for the past 4 seasons from ProBasketballReference.com, cleaned up the data a bit, and appended it into a single worksheet.|
Next, I filtered out all Home games and all games Harden was inactive or DNP.
Poor Performances were determined by 6 stats: Points, FG%, 3PT%, FT%, Assists and Turnovers. For each of these stats I compared Harden's overall season-average to the city-specific average. I identified 2 categories of poor performances:
1.) Sub-Par - Harden performed WORSE than season average, and
2.) Very Sub-Par - Harden performed 20%+ WORSE than season average.
I analyzed his poor performances across each of the NBA's 28 different cities (28 citiese: did not look at home games so no Houston, there are 2 teams in LA, and I distinguished between Brooklyn and NYC).
City Strip Club Rating was determined by the average google review rating for the first 10 strip clubs in each city based on the google search "[CITY] Strip Clubs" (e.g., "Detroit Strip clubs"). Yes, this did involve me making like 30 searches for strip clubs on my work cpu..
Finally, I put the City Strip Club Rating into the pivoted game logs, performed a regression analysis and visualized it into charts.
|Conclusion:||I have proven to a statistically significant degree that James Harden's game performance declines in cities with higher rated strip clubs. |
Correlation Coefficient - r - (between avg strip club rating and total # of sub-par games) = .4575
For the nature of the subject matter, this would be considererd a moderate-to-strong correlation.
Coefficient of Determination - r2 - (between avg strip club rating and total # of sub-par games) = .21
This means that James Harden's box score is 20% predictable based on the quality of a city's strip clubs
Other interesting facts:
1.) Harden's best performance comes in city with the worst strip clubs - Toronto
2.) Harden's worst performance comes in city with the best strip clubs - Miami
3.) Salt Lake city has the 3rd-ranked strip clubs of all NBA cities lol