player_data <-read.csv('C:/Users/rohan/OneDrive/Desktop/INTRO TO STATISTICS IN R/DATA SETS/Datasets/Data/Nba_all_seasons_1996_2021.csv')
summary(player_data)## X player_name team_abbreviation age
## Min. : 0 Length:12305 Length:12305 Min. :18.00
## 1st Qu.: 3076 Class :character Class :character 1st Qu.:24.00
## Median : 6152 Mode :character Mode :character Median :26.00
## Mean : 6152 Mean :27.08
## 3rd Qu.: 9228 3rd Qu.:30.00
## Max. :12304 Max. :44.00
## player_height player_weight college country
## Min. :160.0 Min. : 60.33 Length:12305 Length:12305
## 1st Qu.:193.0 1st Qu.: 90.72 Class :character Class :character
## Median :200.7 Median : 99.79 Mode :character Mode :character
## Mean :200.6 Mean :100.37
## 3rd Qu.:208.3 3rd Qu.:108.86
## Max. :231.1 Max. :163.29
## draft_year draft_round draft_number gp
## Length:12305 Length:12305 Length:12305 Min. : 1.00
## Class :character Class :character Class :character 1st Qu.:31.00
## Mode :character Mode :character Mode :character Median :57.00
## Mean :51.29
## 3rd Qu.:73.00
## Max. :85.00
## pts reb ast net_rating
## Min. : 0.000 Min. : 0.000 Min. : 0.000 Min. :-250.000
## 1st Qu.: 3.600 1st Qu.: 1.800 1st Qu.: 0.600 1st Qu.: -6.400
## Median : 6.700 Median : 3.000 Median : 1.200 Median : -1.300
## Mean : 8.173 Mean : 3.559 Mean : 1.814 Mean : -2.256
## 3rd Qu.:11.500 3rd Qu.: 4.700 3rd Qu.: 2.400 3rd Qu.: 3.200
## Max. :36.100 Max. :16.300 Max. :11.700 Max. : 300.000
## oreb_pct dreb_pct usg_pct ts_pct
## Min. :0.00000 Min. :0.000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.02100 1st Qu.:0.096 1st Qu.:0.1490 1st Qu.:0.4800
## Median :0.04100 Median :0.131 Median :0.1810 Median :0.5240
## Mean :0.05447 Mean :0.141 Mean :0.1849 Mean :0.5111
## 3rd Qu.:0.08400 3rd Qu.:0.180 3rd Qu.:0.2170 3rd Qu.:0.5610
## Max. :1.00000 Max. :1.000 Max. :1.0000 Max. :1.5000
## ast_pct season
## Min. :0.0000 Length:12305
## 1st Qu.:0.0660 Class :character
## Median :0.1030 Mode :character
## Mean :0.1314
## 3rd Qu.:0.1780
## Max. :1.0000# Load necessary libraries
library(glmnet)## Warning: package 'glmnet' was built under R version 4.3.2## Loading required package: Matrix## Loaded glmnet 4.1-8library(caret)## Warning: package 'caret' was built under R version 4.3.2## Loading required package: ggplot2## Loading required package: latticelibrary(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union# Choose response variable and explanatory variables
response_variable <- "pts"
# Select relevant explanatory variables
explanatory_variables <- c("age", "player_height", "player_weight", "gp", "reb", "ast", "usg_pct", "ts_pct")
# Subset the dataset with selected variables
subset_data <- player_data %>% select(response_variable, explanatory_variables)## Warning: Using an external vector in selections was deprecated in tidyselect 1.1.0.
## ℹ Please use `all_of()` or `any_of()` instead.
## # Was:
## data %>% select(response_variable)
##
## # Now:
## data %>% select(all_of(response_variable))
##
## See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.## Warning: Using an external vector in selections was deprecated in tidyselect 1.1.0.
## ℹ Please use `all_of()` or `any_of()` instead.
## # Was:
## data %>% select(explanatory_variables)
##
## # Now:
## data %>% select(all_of(explanatory_variables))
##
## See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.# Check for missing values and remove rows with missing data if necessary
subset_data <- na.omit(subset_data)
# Check if there are at least 2 data points for the response variable
if (nrow(subset_data) < 2) {
stop("Insufficient data points for the response variable.")
}
# Split the data into training and testing sets
set.seed(123) # for reproducibility
train_index <- createDataPartition(subset_data[[response_variable]], p = 0.8, list = FALSE)
train_data <- subset_data[train_index, ]
test_data <- subset_data[-train_index, ]
# Build the generalized linear model (GLM)
glm_model <- glm(paste(response_variable, "~ ."), data = train_data, family = gaussian)
# Make predictions on the test set
predictions <- predict(glm_model, newdata = test_data)
# Evaluate the model performance (you can use appropriate metrics based on your context)
mse <- mean((test_data[[response_variable]] - predictions)^2)
print(paste("Mean Squared Error (MSE):", mse))## [1] "Mean Squared Error (MSE): 7.06216455532447"In order to select explanatory variables, variables such as age, player height, player weight, games played, rebounds, assists, usage percentage, and true shooting percentage are chosen. “Pts” (points scored) is the response variable. Depending on your analysis objectives, you can change the response_variable and explanatory_variables.
The numerical columns from the dataset that we wish to include in the correlation matrix are first specified. The dataset is then subset so that only these numerical columns remain. Lastly, we calculate the correlation coefficients between these numerical variables using the cor() function, which allows us to create the correlation matrix.
correlation_matrix <- cor(subset_data[explanatory_variables])
print("Correlation Matrix:")## [1] "Correlation Matrix:"print(correlation_matrix)## age player_height player_weight gp reb
## age 1.000000000 -0.008954385 0.05854449 0.055217522 0.03562267
## player_height -0.008954385 1.000000000 0.82542135 0.002202118 0.42327549
## player_weight 0.058544493 0.825421354 1.00000000 0.019651800 0.43792982
## gp 0.055217522 0.002202118 0.01965180 1.000000000 0.47094829
## reb 0.035622669 0.423275490 0.43792982 0.470948288 1.00000000
## ast 0.090208519 -0.449033300 -0.37878377 0.385836033 0.24101766
## usg_pct -0.119416223 -0.104713789 -0.06738550 0.146806004 0.23055212
## ts_pct 0.025175472 0.072154033 0.06814165 0.375816016 0.31456949
## ast usg_pct ts_pct
## age 0.09020852 -0.1194162 0.02517547
## player_height -0.44903330 -0.1047138 0.07215403
## player_weight -0.37878377 -0.0673855 0.06814165
## gp 0.38583603 0.1468060 0.37581602
## reb 0.24101766 0.2305521 0.31456949
## ast 1.00000000 0.3925326 0.17608334
## usg_pct 0.39253260 1.0000000 0.12336769
## ts_pct 0.17608334 0.1233677 1.00000000# Assuming you have already loaded the dataset
# nba_data <- read.csv("path/to/your/dataset.csv")
# Fit a logistic regression model (example using 'draft_round' as the outcome variable)
logistic_model <- glm(draft_round == "1" ~ pts, data = player_data, family = binomial)
# Print the coefficients
print(summary(logistic_model))##
## Call:
## glm(formula = draft_round == "1" ~ pts, family = binomial, data = player_data)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.938031 0.035436 -26.47 <2e-16 ***
## pts 0.165192 0.004324 38.20 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16783 on 12304 degrees of freedom
## Residual deviance: 14808 on 12303 degrees of freedom
## AIC: 14812
##
## Number of Fisher Scoring iterations: 4# Interpret the coefficient for 'pts'
pts_coefficient <- coef(logistic_model)["pts"]
print(paste("Coefficient for pts:", pts_coefficient))## [1] "Coefficient for pts: 0.165191797078324"Based on a player’s points per game (pts), we are forecasting whether or not they will be selected in the first round (draft_round == “1”). Holding all other factors constant, the coefficient for points will show how the log-odds of being drafted in the first round alter with each extra point a player scores.
This coefficient’s meaning varies depending on the logistic regression model and dataset you are using. A player’s likelihood of being selected in the first round increases with their point total in a game, according to a positive coefficient; conversely, a negative coefficient would imply the opposite. The coefficient’s magnitude represents the strength of the correlation between points scored per game and the chance of being selected in the first round.