About Me
Hi! I’m Chris, a data scientist and former math teacher.I specialize in using data visualization, predictive modeling, and analytics to drive decisions that make a positive impact on society.With a passion for solving complex problems, I turn data into actionable insights that empower organizations and communities.
Skills
Python | R | Tableau | Power BI | SQL | Excel | InDesign | Git
Data Analysis 2+ years
Data Visualization 2+ years
Data Reporting 1+ years
Predictive Modeling 1+ years
Featured Projects
PYTHON | MACHINE LEARNING
Traffic Crash Prediction
Machine learning classification project in Python analyzing Chicago Crash Data to predict key contributors of fatal crashes, and provide 3 data-driven recommendations.
R & TABLEAU | REGRESSION ANALYSIS
Gun Violence & Education Dashboard
Conducted hypothesis test and regression analysis to explore the relationship between juvenile gun deaths and high school graduation rates across Chicago ZIP codes. Created a Tableau dashboard to visualize the results.
PYTHON | MACHINE LEARNING
Water Well Classification
Implemented a machine learning model to classify water wells using geospatial data and predictive modeling techniques.
EXCEL & CANVA | DATA ANALYSIS
ACT Program Evaluation Report
Created a comprehensive data report analyzing 8 years of high school data, measuring student growth and curriculum effectiveness.
TABLEAU | DATA VISUALIZATION
Trees & Wealth Dashboard
Created Tableau dashboard to explore the relationship between wealth & trees in New York City including a horizontal bar chart, stacked bar charts, bubble chart, & interactive maps.
Professional Experience
DATA/POLICY ANALYST - Cook County Office of Research, Operations & Innovation
VISION ZERO COMMUNITY AMBASSADOR - City of Philadelphia
LEAD STEM CONTENT SPECIALIST - TEACH, LLC.
ARTIFICIAL INTELLIGENCE/MACHINE LEARNING INSTRUCTOR - Kode with Klossy
HIGH SCHOOL MATH TEACHER - School District of Philadelphia
SUPPLY CHAIN & OPERATIONS INTERN - Laborocity
Education
Data, Economics, and Public Policy Certificate, MITx MicroMasters
M.S. Education, University of Pennsylvania
Data Science Certificate, Flatiron School Bootcamp
B.S. Mathematics, University of Pittsburgh
Recent Blog Posts
STARTING A DATA CAREER: WHAT I LEARNED IN MY FIRST 90 DAYS
Key takeaways from my first quarter of my applied data fellowship
HOW TO USE DATA STORYTELLING TO CREATE EFFECTIVE AND IMPACTFUL VISUALS
Key data visualization strategies for effective storytelling
FROM TEACHING TO TECH: MY CAREER TRANSITION INTO DATA SCIENCE
Read about my journey from teaching to data science
VISUALIZATION ESSENTIALS WITH MATPLOTLIB AND SEABORN
Learn the basics of creating data visualizations in Python
INTRO TO HYPOTHESIS TESTING IN PYTHON
Beginner’s guide to using Python to conduct and interpret hypothesis tests
Python | Machine Learning
Traffic Crash Prediction
OVERVIEW:This project analyzes Chicago’s crash data to identify factors influencing severe crashes, aiming to support Vision Zero initiatives and provide actionable insights for policymakers.DATA: Over 900k records on crash incidents, 1.98 million on individuals, and 1.84 million on vehicles.GOALS: Develop an inferential model to predict and identify factors leading to serious crashes, addressing challenges like class imbalance.MODELING STEPS:
1. Data Cleaning: Handled missing values, duplicates, outliers, and ensured proper formatting.
2. Data Splitting: Prepared training, validation, and test sets to prevent data leakage.
3. Baseline Models: Built logistic regression and decision tree models for initial comparisons.
4. Addressing Class Imbalance: Applied SMOTE, class weighting, and their combinations to address imbalances.
5. Hyperparameter Tuning: Used GridSearchCV to find optimal parameters, optimizing for PR AUC.
6. Model Selection: Chose the decision tree for its interpretability and strong performance.
7. Model Evaluation: Focused on evaluating feature importances from the decision tree to gain insights into key predictors.
- INSIGHT #1: Overrepresented factors in severe crashes include airbag deployment, light condition, and males.
- INSIGHT #2: The top contributing factors identified in severe crashes are airbag deployment, males, and seasonality. The feature importance output from the decision tree model highlights these variables as having the greatest influence on crash severity.
RECOMMENDATIONS:
1. Male Injury Trends:
- Conduct further research to analyze trends in male driver involvement in serious crashes.
- Develop targeted safety campaigns focused on behaviors like speeding and aggressive driving.2. Vehicle Size and Speed Regulations:
- Implement city-specific tiered fees for heavier vehicles, using revenue to fund pedestrian and cyclist safety measures.3. Annual Airbag Inspections:
- Promote policies requiring annual airbag inspections to ensure functionality and compliance with safety standards.
The full project can be viewed on my github page: here
R & TABLEAU | REGRESSION ANALYSIS
Gun Violence & Education Dashboard
OVERVIEW:
Explored the relationship between youth gun deaths and high school graduation rates across Chicago ZIP codes using publicly available data (2018-2021). Conducted hypothesis test, and regression analysis, as well as created data visualizations, revealing a statistically significant inverse relationship between graduation rates and juvenile gun deaths.DATA:
Mortality Data: Juvenile (under 18) gun-related deaths in Chicago (2018–2021).
Education Data: 4-year high school graduation rates by ZIP code (2018–2021).
GOALS:
Explore the relationship between youth gun deaths and high school graduation rates by:
1. Conducting hypothesis test to determine the statistical significance of the relationship.
2. Performing regression analysis to quantify the strength and direction of this relationship.
3. Visualizing the data using maps and scatterplots to illustrate the relationship both geographically and quantitatively.
KEY INSIGHTS:
The relationship between graduation rates and juvenile gun deaths is inverse—as graduation rates increase, juvenile gun deaths decrease.
The relationship is statistically significant (p = 0.0216, p < 0.05).
For every 1% increase in the graduation rate, there is an associated decrease of 0.349 juvenile gun deaths per 100K youth.
The full project can be viewed on my github page: here
Python | Machine Learning
Water Well Classification
OVERVIEW:
This project uses machine learning to assist an NGO in identifying damaged water wells across Tanzania.DATA: The dataset includes 59k water well records with 41 features, such as location, water source, and functionality status.GOAL: Build a machine learning model that classifies wells as functional or in need of repair, with a focus on maximizing recall to minimize missed repair opportunities.MODELING STEPS:
1. Data cleaning and preprocessing, including handling null values, reclassification of target variable, and feature engineering.
2. Trained multiple models, including logistic regression, decision trees, and random forests.
3. Evaluated performance using recall to minimize false negatives, with an emphasis on regions most in need of repairs.
INSIGHTS:
Areas like Kigoma and Dar es Salaam are high priority for repairs due to a high density of broken wells and low access to functional ones.
RECOMMENDATIONS:
1. Prioritize regions identified for high repair needs.
2. Deploy the initial Random Forest model to predict repair needs, as it offers the best recall despite some overfitting.
3. Refine the model further with better feature selection and hyperparameter tuning to improve performance and reduce overfitting.
The full project can be viewed on my github page: here
Excel & Canva | Data Analysis
ACT PROGRAM EVALUATION Report
OVERVIEW:
Developed a 5-page data report analyzing the effectiveness of the ACT Prep program over 8 years, summarizing key trends, improvements in scores, and national comparisons.DATA:
Analyzed data from 8 cohorts, serving 156 students with an average enrollment of 19.5 students per cohort.INSIGHTS
93% of students achieved growth
On average, students’ national percentiles increased by 17 points, rising from the 22nd to the 39th percentile during the ACT Prep class
The full report can be viewed: here
Tableau | Data Visualization
Trees & Wealth Dashboard
OVERVIEW:
Developed an interactive Tableau dashboard to explore the relationship between wealth and trees in New York City.DATA:
Income data across NYC’s 5 boroughs.
Tree census data from over 680k trees
The full dashboard can be viewed on my Tableau Public page: here