In [163]:
import matplotlib.pyplot as plt
import numpy as np
from sklearn import linear_model
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
import pandas as pd
import seaborn as sns
In [145]:
players_df = pd.read_csv('players.csv')
salaries_df = pd.read_csv('salaries_1985to2018.csv')
data_df = pd.merge(players_df, salaries_df, on= 'player_id')
In [147]:
data_df.head()
Out[147]:
index_x player_id birthDate birthPlace career_AST career_FG% career_FG3% career_FT% career_G career_PER ... position shoots weight index_y league salary season season_end season_start team
0 0 abdelal01 24-Jun-68 Cairo, Egypt 0.3 50.2 0 70.1 256 13 ... Power Forward Right 240lb 0 NBA 395000 1990-91 1991 1990 Portland Trail Blazers
1 0 abdelal01 24-Jun-68 Cairo, Egypt 0.3 50.2 0 70.1 256 13 ... Power Forward Right 240lb 1 NBA 494000 1991-92 1992 1991 Portland Trail Blazers
2 0 abdelal01 24-Jun-68 Cairo, Egypt 0.3 50.2 0 70.1 256 13 ... Power Forward Right 240lb 2 NBA 500000 1992-93 1993 1992 Boston Celtics
3 0 abdelal01 24-Jun-68 Cairo, Egypt 0.3 50.2 0 70.1 256 13 ... Power Forward Right 240lb 3 NBA 805000 1993-94 1994 1993 Boston Celtics
4 0 abdelal01 24-Jun-68 Cairo, Egypt 0.3 50.2 0 70.1 256 13 ... Power Forward Right 240lb 4 NBA 650000 1994-95 1995 1994 Sacramento Kings

5 rows × 32 columns

In [149]:
data_df.columns
Out[149]:
Index(['index_x', 'player_id', 'birthDate', 'birthPlace', 'career_AST',
       'career_FG%', 'career_FG3%', 'career_FT%', 'career_G', 'career_PER',
       'career_PTS', 'career_TRB', 'career_WS', 'career_eFG%', 'college',
       'draft_pick', 'draft_round', 'draft_team', 'draft_year', 'height',
       'highSchool', 'name', 'position', 'shoots', 'weight', 'index_y',
       'league', 'salary', 'season', 'season_end', 'season_start', 'team'],
      dtype='object')
In [151]:
data_df.isna().sum()
Out[151]:
index_x            0
player_id          0
birthDate          0
birthPlace         0
career_AST         0
career_FG%         0
career_FG3%        0
career_FT%         0
career_G           0
career_PER         0
career_PTS         0
career_TRB         0
career_WS          0
career_eFG%        0
college         1636
draft_pick      1902
draft_round     1902
draft_team      1902
draft_year      1902
height             0
highSchool       989
name               0
position           0
shoots             0
weight             0
index_y            0
league             0
salary             0
season             0
season_end         0
season_start       0
team               4
dtype: int64
In [152]:
cleaned_data_df = data_df.drop(['college', 'draft_pick', 'draft_round', 'draft_team', 'draft_year', 'highSchool'], axis = 1)
In [155]:
cleaned_data_df.isna().sum()
Out[155]:
index_x         0
player_id       0
birthDate       0
birthPlace      0
career_AST      0
career_FG%      0
career_FG3%     0
career_FT%      0
career_G        0
career_PER      0
career_PTS      0
career_TRB      0
career_WS       0
career_eFG%     0
height          0
name            0
position        0
shoots          0
weight          0
index_y         0
league          0
salary          0
season          0
season_end      0
season_start    0
team            4
dtype: int64
In [157]:
sns.scatterplot(x='career_PTS', y='salary', data=data_df)
Out[157]:
<Axes: xlabel='career_PTS', ylabel='salary'>
No description has been provided for this image
In [158]:
sns.scatterplot(x='career_G', y='salary', data=data_df)
Out[158]:
<Axes: xlabel='career_G', ylabel='salary'>
No description has been provided for this image
In [171]:
X = cleaned_data_df[['career_AST','career_TRB','career_PTS','career_G']]
y = cleaned_data_df['salary']
In [175]:
from sklearn.preprocessing import StandardScaler

scaled = StandardScaler().fit_transform(x)

X_scaled = pd.DataFrame(scaled, columns = X.columns, index = X.index)
In [177]:
X_scaled.head()
Out[177]:
career_AST career_TRB career_PTS career_G
0 -0.998645 -0.274361 -0.663454 -0.962424
1 -0.998645 -0.274361 -0.663454 -0.962424
2 -0.998645 -0.274361 -0.663454 -0.962424
3 -0.998645 -0.274361 -0.663454 -0.962424
4 -0.998645 -0.274361 -0.663454 -0.962424
In [161]:
y.head()
Out[161]:
0    395000
1    494000
2    500000
3    805000
4    650000
Name: salary, dtype: int64
In [ ]:
In [179]:
X_train, X_test, y_train, y_test = train_test_split(X_scaled,y, test_size = 0.3, random_state = 7)
In [181]:
lr = linear_model.LinearRegression()

lr_model = lr.fit(X_train,y_train)

lr_model.score(X_test,y_test)
Out[181]:
0.25364217200690564
In [183]:
y_pred = lr_model.predict(X_test)
In [195]:
from sklearn import metrics

mae = metrics.mean_absolute_error(y_test,y_pred)
mse = metrics.mean_squared_error(y_test, y_pred)
rmse = metrics.mean_squared_error(y_test, y_pred, squared = False)

/opt/anaconda3/lib/python3.12/site-packages/sklearn/metrics/_regression.py:483: FutureWarning: 'squared' is deprecated in version 1.4 and will be removed in 1.6. To calculate the root mean squared error, use the function'root_mean_squared_error'.
  warnings.warn(
In [197]:
rmse
Out[197]:
3694693.4692689087
In [125]:
reg.coef_
Out[125]:
array([119765.21651724, 376500.82994602, 255351.87623666,    379.91311393])
In [127]:
reg.score(x,y)
Out[127]:
0.2630190753541878
In [129]:
coefficients = pd.concat([pd.DataFrame(x.columns), pd.DataFrame(np.transpose(reg.coef_))], axis = 1)
In [131]:
coefficients
Out[131]:
0 0
0 career_AST 119765.216517
1 career_TRB 376500.829946
2 career_PTS 255351.876237
3 career_G 379.913114
In [ ]:
from sklearn.feature_selection import SelectKBest, f_regression

fs = SelectKBest(score_func = f_regression, k = 2)

best = fs.fit_transform(X_scaled,y)

features = X_scaled.columns 

best_features = features[fs.get_support()]