MA 346 Final Project

Matt Bacon and Danny Ramones

import pandas as pd import matplotlib.pyplot as plt tackles = pd.read_csv( 'tackles.csv' ) tackles

players = pd.read_csv( 'players.csv' ) players

# Merge datasets on 'nflId' merged_data = pd.merge(players, tackles, on='nflId') merged_data

Graph 1

# Calculate total tackles for each player total_tackles_by_player = merged_data.groupby(['nflId', 'displayName', 'collegeName'])['tackle'].sum().reset_index() # Find the top 30 players based on total tackles top_players = total_tackles_by_player.nlargest(30, 'tackle') # Increase figure size and adjust bar width for better spacing fig, ax = plt.subplots(figsize=(12, 6)) bar_width = 0.8 # Plot a bar chart for total tackles by college for the top 30 players plt.bar(top_players['displayName'] + ' - ' + top_players['collegeName'], top_players['tackle'], width=bar_width, color='skyblue') plt.title('Total Number of Tackles for Top 30 Players') plt.xlabel('Player Name - College Name') plt.ylabel('Total Number of Tackles') plt.xticks(rotation=45, ha='right') # Rotate x-axis labels for better readability plt.tight_layout() # Adjust layout to prevent clipping plt.show()

Graph 2

# Calculate total tackles for each college total_tackles_by_college = merged_data.groupby('collegeName')['tackle'].sum().reset_index() # Find the top 10 colleges based on total tackles top_colleges = total_tackles_by_college.nlargest(10, 'tackle') # Plot a bar chart for total tackles by college for the top 10 colleges plt.bar(top_colleges['collegeName'], top_colleges['tackle'], color='skyblue') plt.title('Total Number of Tackles for Top 10 Colleges') plt.xlabel('College Name') plt.ylabel('Total Number of Tackles') plt.xticks(rotation=45, ha='right') # Rotate x-axis labels for better readability plt.show()

Statistical Analysis

from scipy.stats import spearmanr # Calculate Spearman correlation coefficient and p-value correlation_coef, p_value = spearmanr(merged_data['collegeName'].astype('category').cat.codes, merged_data['tackle']) # Print the results print(f"Spearman's Rank Correlation Coefficient: {correlation_coef}") print(f"P-value: {p_value}") # Check for significance (using a common significance level of 0.05) if p_value < 0.05: print("The correlation between college and tackles is statistically significant.") else: print("There is no significant correlation between college and tackles.")

Overall Analysis

Project Summary and Goals:

The primary goal of this project was to investigate the correlation between the college attended by NFL players and the number of tackles they make in the league. Our objective was to provide insights into whether a player's college choice is a significant factor in predicting their performance in terms of tackles.

Data Selection and Preparation:

The data used in this project originated from two datasets we downloaded from Kaggle: 'players' and 'tackles.' The 'players' dataset included information such as 'nflId,' 'collegeName,' and 'displayName,' while the 'tackles' dataset contained 'nflId' and 'tackles.' Merging these datasets based on 'nflId' allowed us to create a comprehensive dataset for analysis.

To ensure the data was suitable for our goals, we performed necessary data cleaning and transformation steps. Notably, we aggregated the 'tackles' dataset to calculate the total number of tackles for each player, providing a more accurate representation of player performance.

Statistical Analysis:

We employed Spearman's rank correlation coefficient to assess the relationship between 'collegeName' (categorical) and the total number of tackles. The analysis did not reveal a statistically significant correlation, indicating that the choice of college is not a reliable predictor of tackles in the NFL.

Top Colleges and Tackles:

Despite the absence of a significant correlation, we identified the top 10 colleges with the most tackles: Alabama, LSU, Ohio State, Washington, Florida, Georgia, Texas, Notre Dame, Penn State, and UCLA. Our analysis considered the implications of this diversity and suggested that player success is influenced by various factors beyond the geographic location or prestige of the college.

Observations:

Diversity in Top Colleges: The list of top colleges with the most tackles is diverse, representing various regions across the United States. This suggests that successful NFL players, in terms of tackles, emerge from colleges nationwide rather than being concentrated in a specific geographic area.

Notable Football Programs: Many of the top colleges are renowned for their strong football programs. Colleges like Alabama, Ohio State, and LSU have consistently produced successful NFL players in various positions, indicating a strong football culture and effective player development.

Varied Playing Styles: The lack of a significant correlation does not necessarily indicate a lack of influence; it may suggest that the style of play or the skills developed in college football programs does not directly translate to the number of tackles a player makes in the NFL. Different colleges may emphasize different aspects of player development.

Potential for Further Investigation: While our analysis does not identify a clear correlation, there may be other factors at play, such as player positions, team strategies, or changes in the game over time. Further investigation into these factors could provide a more nuanced understanding of the relationship between college attended and NFL tackles.

Conclusion:

In conclusion, this project showcases the application of statistical analysis to real-world data, highlighting the nuanced relationship between college attended and player performance in the NFL. The findings offer valuable insights for decision-makers in both sports analytics and broader data-driven contexts.

References:

https://www.kaggle.com/competitions/nfl-big-data-bowl-2024/data https://chat.openai.com/