This page is collection of my data science projects. The posts are organized by projects. Each set of posts follow a similar pattern. First I present the algorithm I used on a toy data set, then present exploratory data analysis and then build predictive model.

Project 1: Data-driven techniques for identifying at-risk students.

In this project, I develop a student intervention system based on demographic, economic and behavioral data collected in students from two different schools. I first convert all the data to categorical variables and apply Multiple Correspondence Analysis (MCA). MCA is a principal component analysis like technique for cases where all the variables are categorical.