My Projects

I collaborated closely with my teammates and engaged in external discussions with Genetic Alliance executives to transform a vast disease database into a practical and insightful representation, aiming to alleviate the challenges associated with cost, time, and the high 95% failure rate in disease research. I designed and implemented a sophisticated algorithm using Python and SQL. This algorithm effectively distilled and categorized a substantial dataset comprising 100 fields and 10,000 rows of information. The outcome was a weighted scoring system that facilitated the pairing of research groups with diseases aligned to their focus. Through this initiative, we sought to streamline research processes, enhance efficiency, and contribute to a more targeted and successful approach in the field of disease research.

While working for the University of Michigan's Language Resource Center, I conducted comprehensive analytics of two key datasets: Film Reservations and Calendar Data. Using the Film Reservation Data, I employed Python's Pandas library for data import and initial exploration, followed by meticulous data cleaning processes. This included course cleaning, handling missing data, and column renaming, ultimately leading to key summary numbers and in-depth insights into enrollment patterns, fair use versus licensed films, correlation analysis, course levels served, and duration and start month analysis. Subsequently, the Calendar/Room Reservation Data underwent a similar rigorous analysis, involving data import, cleaning, and exploration for multiple rooms. The key summary numbers and visualizations provided detailed information on total reservations, average reservation length, room-to-room comparisons, day-of-week analysis, and a heatmap illustrating room popularity. Collectively, the analysis of both datasets offers valuable insights into course enrollment, reservation distribution, and film usage trends, providing a foundation for informed decision-making and resource allocation in future semesters.

I extracted data from online databases and employed data integration techniques, statistical quantitative analysis, and data exploration libraries in Python to delve into 7,520,000 rows of Chicago crime data spanning the last two decades. This extensive analysis involved a meticulous examination of patterns and trends within the dataset. Subsequently, I applied data visualization techniques to present the findings in a comprehensible manner, unveiling key insights into the dynamics and patterns of crime over the 20-year period. This process not only facilitated a deep understanding of the dataset but also allowed for the identification of significant trends that contribute to a more informed perspective on the evolving nature of crime in Chicago over time.

In my involvement with the ACLED (Armed Conflict Location & Event Data) project, I focused on the disaggregated data collection, analysis, and crisis mapping initiative. This project encompasses a vast dataset detailing political violence and disorder events globally. The dataset I worked with specifically provided a non-aggregated summary of every event documented worldwide in 2022, consisting of 10,000 randomly selected data points. To efficiently manage and analyze this extensive dataset, I leveraged Pandas and Jupyter notebook. The initial step involved importing the large and intricate data as a CSV file into a Pandas dataframe, followed by an intensive data cleaning process to address inconsistencies and ensure accuracy. For enhanced analysis, I extracted key information such as countries, regions, event types, and disorder types from the data. Subsequently, I converted these details into integer IDs and manually inputted this information into DataGrip for streamlined database management. The final step involved importing the cleaned events table back into DataGrip for further exploration and analysis. This comprehensive process of importing, normalizing, and cleaning the ACLED dataset was crucial for preparing the data for subsequent in-depth analyses and mapping activities within the context of global political violence and disorder events.

In my Piazza project on Machine Learning, I focused on implementing natural language processing and machine learning techniques to automatically identify the subject of posts within a question/learning platform. The project emphasized the mastery of Function Objects and Recursion, reinforcing skills acquired from previous projects such as Container ADTs, Dynamic Memory, The Big Three, Linked Lists, and Iterators. Throughout this endeavor, I gained hands-on experience in employing recursion, working with binary trees, utilizing templates, implementing comparators, and utilizing the map data structure. The ultimate goal was to create a program capable of parsing and comprehending Piazza posts, enabling it to accurately discern the subject matter. An outcome of this project involved the program successfully interpreting a Piazza post and determining its relevance to a specific subject, demonstrating the practical application of natural language processing and machine learning in solving challenges related to information extraction and classification.