Applied Statistics
Through these projects, I demonstrated strong statistical analysis and data modeling skills using historical basketball data to derive actionable insights. I applied descriptive statistics (mean, median, variance, standard deviation) to summarize team performance and evaluated confidence intervals to assess player skill levels. I also performed multiple hypothesis tests—including tests for population mean, population proportion, and differences between two population means—to evaluate scoring trends and relative skill comparisons between teams.
​
Additionally, I utilized Python-based regression modeling, including simple linear regression, multiple regression, and OLS models, to quantify relationships between variables such as total wins, points scored, relative skill, and point differentials. Correlation analysis and regression coefficients allowed me to predict team outcomes and assess the impact of different performance metrics. These projects collectively highlight my ability to apply data visualization, statistical reasoning, and predictive modeling to real-world datasets for informed decision-making.
Project One
This project analyzed historical basketball data to evaluate the Sacramento Kings’ performance during the 2013–2015 seasons. Using descriptive statistics such as mean, median, variance, and standard deviation, along with confidence intervals, the Kings’ scoring trends were compared to the Chicago Bulls’ 1996–1998 seasons. Data visualization techniques, including scatter plots and box plots, highlighted scoring patterns, averages, and variability, while comparisons of home and away games revealed minor differences in performance.
​
The analysis found that while the Kings’ scoring slightly increased over the period and player skill levels were higher than those of the Bulls, the 20-year gap between the teams limited actionable insights. The project concluded that tracking data from teams within the same timeframe, particularly home versus away performance, would provide more relevant insights for improving team strategies and performance in upcoming seasons.
Project Two
This project analyzed historical data of the Sacramento Kings (2013–2015) to evaluate team performance and identify strategies to improve future outcomes. Using hypothesis tests for population mean, population proportion, and differences between two population means, key metrics such as average points per game and relative skill levels were examined. Results showed that the Kings had an average Elo score of 1411.04 and scored an average of 100.69 points per game, with a 60% increased chance of winning when scoring at least 102 points. Statistical analysis also revealed that the Kings’ skill level was significantly higher than the Chicago Bulls from 1996–1998, though the large gap in years limited the relevance of direct comparisons.
​
The findings indicate that monitoring contemporary data—particularly scoring trends and skill levels—can guide actionable strategies for the upcoming season. By focusing on incremental improvements, such as scoring slightly more per game, the Kings can potentially increase their win ratio. The project highlights the importance of applying rigorous statistical methods to sports analytics to provide meaningful insights for team development and decision-making.
Project Three
​
This project analyzed historical basketball data using Python to identify patterns that predict the Sacramento Kings’ total wins in a season. Scatterplots and correlation analyses showed strong positive relationships between wins and average relative skill (Pearson correlation = 0.907) and a moderate positive relationship with average points scored (Pearson correlation = 0.478). Simple and multiple OLS regression models were used to quantify these relationships, confirming that both relative skill and points scored significantly impact total wins.
​
The final multiple regression model incorporated average points, relative skill, points differential, and ELO differential to predict total wins, accounting for 87% of the variance. Predictions from the model demonstrate how maintaining high points and skill levels, along with outperforming opponents, can maximize the team’s success. This analysis provides actionable insights for strategic planning, goal-setting, and tracking performance across the season to improve competitive outcomes.