I built a machine learning system classifying breast cancer cases as benign or malignant with 97% accuracy from cellular measurements. After testing multiple algorithms, I found 10 key features maintain performance, making it practical for clinical diagnosis assistance.

In this project, I delved into the fascinating world of global cancer patient data from 2015 to 2024. I leveraged Python for data cleaning and extraction, and then brought my findings to life with interactive visualizations in Flourish Studio.

I hopped on a forensic analysis using SQL. I also had help from a colleague of mine, Ojibara Farida. We were able to determine the killer, and prime suspects from the data provided, although not explicitly defined, I believe our findings should be close enough. Click on the link below to check out how we approached the problem.

My aim was to build a classification model to predict previous gestational diabetes based on some features. I automated the EDA process using ydataprofiling library. I also used the logistic regression model and some other classification models. The accuracy was below 53%, and most probably due to the data.

In this project, I built a multiple linear regression model predicting the yearly amount spent of customers who bought clothes online using an app and a website using multiple and single features respectively using the F-Regression statistic for feature selection.