hamsini1692 / wids-datathon2024

Predict a model to find out if a patient received metastatic cancer diagnosis screening within 90 days

WIDS-Datathon2024

Predict a model to find out if a patient received metastatic cancer diagnosis screening within 90 days

This repository documents my participation in the WiDS Worldwide datathon, aimed at tackling healthcare inequity through the lens of breast cancer treatment and diagnosis. Sponsored by Gilead Sciences and leveraging a rich dataset from Health Verity enriched with geo-demographic and toxicology data, the challenge focused on predicting metastatic cancer diagnosis within 90 days of screening. My approach integrated advanced data science techniques, including machine learning models and exploratory data analysis, to uncover disparities in healthcare access and their impact on timely treatment. The urgency of this datathon is underscored by the focus on metastatic Triple-Negative Breast Cancer (TNBC), a highly aggressive form of cancer that necessitates swift, effective treatment. The disparity in treatment wait times serves as an indicator of broader inequities in healthcare access. Thus, the datathon's dual objectives emerged: to explore the relationship between patient demographics and the timeliness of treatment, and to assess the impact of environmental hazards on diagnosis and treatment efficacy.