DEA.Rmd

```{r  install package}
install.packages("dplyr")
```

```{r load library}
library(readxl)
library(Benchmarking)
library(ggplot2)
library(dplyr)
```

```{r read data}
defenders_pam_df <- read_excel("C:/Users/marce/OneDrive/Documents/Sports Interactive/Football Manager 2023/Data 25-26/Summer Transfer 2025.xlsx", sheet = "PAM - Defending Duty")

defenders_pam_df
```

```{r function for dea output oriented}
run_dea_out <- function(inputs, outputs) {
    dea_result <- dea(X = inputs, Y = outputs, RTS = "crs", ORIENTATION = "out")
    return(dea_result)
}
```

```{r function for dea input oriented}
run_dea_in <- function(inputs, outputs) {
    dea_result <- dea(X = inputs, Y = outputs, RTS = "crs", ORIENTATION = "in")
    return(dea_result)
}
```


```{r set inputs and outputs}
inputs_def_df<- as.matrix(defenders_pam_df[, c("Poss Lost/90", "Salary", "Mins")])
outputs_def_df <- as.matrix(defenders_pam_df[, c("Aer A/90", "Pr passes/90", "Press Completion Ratio/90", "Poss Won/90", "K Tck/90", "K Hdrs/90", "Dist/90", "Sprints/90", "Hdr %", "Pas %", "Tck R", "Defensive Action/90")])
```

```{r dea result output}
dea_results_out_df <- run_dea_out(inputs_def_df, outputs_def_df)
summary(dea_results_out_df)
```

```{r dea result input}
dea_results_in_df <- run_dea_in(inputs_def_df, outputs_def_df)
summary(dea_results_in_df)
```

```{r add efficiency column in df}
efficiency_scores_out <- as.data.frame(dea_results_out_df$eff)
efficiency_scores_in <- as.data.frame(dea_results_in_df$eff)

colnames(efficiency_scores_out) <- "Efficiency Output Oriented"
colnames(efficiency_scores_in) <- "Efficiency Input Oriented"

defenders_pam_df <- cbind(defenders_pam_df, efficiency_scores_out, efficiency_scores_in)

tail(defenders_pam_df)
```
```{r slice df for salford def}
defenders_salford_df <- tail(defenders_pam_df, 11)

print(defenders_salford_df)
```

```{r function for efficiency plot output oriented dea}
plot_efficiency_chart <- function(df, eff_column, chart_title) {
  df_sorted <- df[order(df[[eff_column]], decreasing = TRUE),]

  plot <- ggplot(df_sorted, aes(x = reorder(Name, df[[eff_column]]), y = df[[eff_column]])) +
    geom_bar(stat = "identity", fill = 'red') +
    geom_text(aes(label = sprintf("%.3f", df[[eff_column]])), hjust = -0.1, size = 2.5) +  
    labs(title = chart_title, 
         x = "Player Name", 
         y = eff_column) +
    coord_flip() +  
    theme_minimal() +  
    theme(axis.text.y = element_text(size = 8), 
          axis.text.x = element_text(size = 10), 
          plot.title = element_text(size = 16))  
}

```

```{r view plot output oriented dea eff for salford def}
def_output_plot <- plot_efficiency_chart(
  df = defenders_salford_df, 
  eff_column = "Efficiency Output Oriented", 
  chart_title = "Salford Defenders Output Oriented Efficiency Scores"
)

print(def_output_plot)
```

```{r view plot input oriented dea eff for salford def}
def_input_plot <- plot_efficiency_chart(
  df = defenders_salford_df, 
  eff_column = "Efficiency Input Oriented", 
  chart_title = "Salford Defenders Input Oriented Efficiency Scores"
)

print(def_input_plot)
```

```{r function for visual efficiency plot on scouted players}
plot_filtered_efficiency_chart <- function(df, eff_column, chart_title) {
  df_filtered <- df %>% 
    filter(.data[[eff_column]] != 1) %>%
    slice(1:(n() - 11))
  df_sorted <- df_filtered[order(df_filtered[[eff_column]], decreasing = TRUE),]
  plot <- ggplot(df_sorted, aes(x = reorder(Name, .data[[eff_column]]), y = .data[[eff_column]])) +
    geom_bar(stat = "identity", fill = 'maroon') +
    geom_text(aes(label = sprintf("%.3f", .data[[eff_column]])), hjust = -0.1, size = 2.5) +  
    labs(title = chart_title, 
         x = "Player Name", 
         y = eff_column) +
    coord_flip() +  
    theme_minimal() + 
    theme(axis.text.y = element_text(size = 8), 
          axis.text.x = element_text(size = 10),  
          plot.title = element_text(size = 16))  
}
```

```{r view plot output oriented for non eff scouted def}
scouted_def_noneff_output_plot <- plot_filtered_efficiency_chart(
  df = defenders_pam_df, 
  eff_column = "Efficiency Output Oriented", 
  chart_title = "Scouted Output Oriented Non-Efficient Defenders Scores"
)
print(scouted_def_noneff_output_plot)
```

```{r view plot input oriented for non eff scouted def}
scouted_def_noneff_input_plot <- plot_filtered_efficiency_chart(
  df = defenders_pam_df, 
  eff_column = "Efficiency Input Oriented", 
  chart_title = "Scouted Input Oriented Non-Efficient Defenders Scores"
)
print(scouted_def_noneff_input_plot)
```

```{r function for slicing df to find eff scouted players}
filter_efficiency <- function(df, eff_column, club) {
  df_filtered <- df %>%
    filter(df[[eff_column]] == 1, df[["Club"]] != club)
  return(df_filtered)
}
```

```{r find efficient output oriented scouted def}
scouted_def_eff_output <- filter_efficiency(
  df = defenders_pam_df, 
  eff_column = "Efficiency Output Oriented",
  club = "Salford"
)

print(scouted_def_eff_output)
```

```{r find efficient input oriented scouted def}
scouted_def_eff_input <- filter_efficiency(
  df = defenders_pam_df, 
  eff_column = "Efficiency Input Oriented", 
  club = "Salford"
)

print(scouted_def_eff_input)
```

```{r function for sensitivity analysis}
sensitivity_analysis <- function(inputs, outputs, orientation) {
  variations <- seq(0.8, 1.2, by = 0.1)
  sensitivity_results <- list()
  
  for (var in variations) {
    modified_inputs <- inputs * var
    modified_outputs <- outputs * var
    dea_result <- dea(X = modified_inputs, Y = modified_outputs, RTS = "crs", ORIENTATION = orientation)
    sensitivity_results[[paste0("Variation_", var)]] <- dea_result$eff
  }
  
  return(sensitivity_results)
}
```

```{r sensitivity analysis output oriented dea}
sensitivity_results_output_df <- sensitivity_analysis(inputs_def_df, outputs_def_df, "out")
sensitivity_results_output_df
```

```{r sensitivity analysis for input oriented dea}
sensitivity_results_input_df <- sensitivity_analysis(inputs_def_df, outputs_def_df, "in")
sensitivity_results_input_df
```