Goals, Leagues and Machine Learning

---

English Premier League is the major football league in England. The project aims at predicting the winner of each fixture every week using statistics about a team's past performance, recent form and home advantage.

• Tools: Python (Pandas, Matplotlib, scikitlearn), SQL


• Date: January to March 2017
• Project Link: Github Link

Close

Digit Recognition using Deep Learning

---

Deep Learning using the H20 Package to classify digits in the MNIST dataset. Even the most basic solution leads to a 96% accuracy in Kaggle.

• Tools: Python, H20 (Deep Learning Framework)


• Date: February 2017
• Project Link: Github Link

Close

West Nile Virus Prediction

---

A project aimed at analyzing data on mosquito traps laid to curtail the spread of the West Nile Virus in Chicago. Along with the weather data, clear patterns emerged about the conditions in which the virus thrives.

A predictive model using Random Forests was developed to indicate the areas in which Chicago health department should concentrate to reduce spending on laying traps.

• Tools: Python (pandas, scikitlearn, seaborn)


• Date: January to April 2017
• Project Link: Github Link

Close

Predicting Loan Defaults for Lending Club

---

Completed this project as a part of the course "Applied Predictive Analytics for Business" at Texas A&M. In this project, we aim at finding a good model which can accurately predict whether a loan applicant would default in the future or not. More importantly, tree based models were explored to supplement the work that has already been done in this area.

An XGboost model proved to the best among the chosen methods under the given assumptions.
Here's the final presentation.

• Tools: R, Machine Learning


• Date: August to December 2016
• Project Link: Report

Close

Other Mini Projects

---

Inferential Statistics

Human Body Temperature - hypothesis testing and statistical significance.
Examining Racial Discrimination - impact of race on the rate of callbacks?
Reducing Hospital Readmissions - statistical analysis to reduce readmissions.

Machine Learning

Customer Segmentation in Python - exploring k-means clustering
Predicting Boston Housing Prices - a mini project in linear regression
Movie Ratings in Rotten Tomatoes - Naive Bayes with NLP
Ghouls, Goblins and Ghosts ... boo! - a fun Kaggle competition.

• Tools: Python (pandas, numpy), Statistics, Machine Learning


• Date: September to December 2016
• Project Link: Github Link

Close

Data for Democracy

---

An active volunteer at Data for Democracy. I am an analyst advocate for the USA Dashboard project and also contribute to the Educational Hub project.

• Tools: Python, Data Collection


• Date: January 2017 till present
• Project Link: Education Hub

Close