Cellular Dysplasia: Predicting Breast Cancer

Last year, breast cancer took the lead in the number of new cancer cases reported, with 2.26 million diagnoses. Although prominent, breast cancer is not as deadly as other cancers like lung, colon, liver, and stomach, mainly because it benefits from early detection. Early detection strategies have been implemented throughout the years and rely on both clinical and self breast exams as well as mammogram screening. If a lump or abnormality is detected, a biopsy will be taken to assess whether the cells are cancerous or not. The biopsy will result in a diagnosis, and the next steps will be dictated from there. It is important to note that earlier-stage cancers can be treated more effectively, which readily contributes to survival. So the goal is always to identify cancer early.

In this project, I will use the Breast Cancer Wisconsin (Diagnostic) Data Set to train a model to predict whether a tumor is benign or malignant given fine needle aspirate biopsies.

Table of Contents

1. Source Data

• Breast Cancer Diagnostic Data Set

2. Documents

• Final Report - This report contains a detailed account of the data wrangling, EDA and modeling performed for this project.
• Presentation - This presentation summarizes the data wrangling, EDA, modeling and conclusions for this project.

3. Notebooks

• Exploration Data Analysis
• Feature Selection
• Machine Learning