Andrea Dorella

Logical Data Analyst skilled in requirement analysis and database management. Self-directed and proactive professional with two years of experience collecting, cleaning and interpreting data sets. Natural problem-solver possessing strong cross-functional understanding of information technology and business processes.

Relevant Project: Graduation Rate Prediction

Rockhurst University - The objective was to employ advanced analytics to enhance graduation rates by identifying students at risk of not graduating within four years.

Accomplishments:

Probability Analysis: Calculated the probability of graduating within four years, providing a quantitative foundation for targeted interventions.
Gini Impurity Index: Evaluated the Overall Gini Impurity Index, gaining insights into the diversity and homogeneity of the dataset.

Gender Impact Assessment: Conducted a thorough analysis to determine the impact of gender on graduation rates, identifying potential gender-based disparities.
Data Visualization: Plotted GPA against SAT scores, enabling a visual exploration of the relationship and distribution of key academic indicators.
Decision Tree Modeling: Implemented recursive partitioning to develop a decision tree, identifying critical predictors influencing graduation outcomes.
Random Forest Implementation: Applied Random Forest for ensemble learning, improving model accuracy and resilience against overfitting.
Policy Recommendations: Pruned the decision tree for enhanced interpretability, leading to strategic policy recommendations for targeted student support.

Outcomes:

Developed a predictive model that aids in proactive identification and support for students at risk of not graduating within the desired timeframe. Enhanced decision-making processes by providing actionable insights derived from statistical analysis and machine learning techniques. Contributed to the establishment of data-driven policies, fostering a culture of continuous improvement in education outcomes.