Grad-school archive
Coursework and side projects from my master’s at Georgia Tech, 2020 – 2022: machine learning, data visualization and a little programming. Kept for reference — my current work is on the home page.
AboutArchive
-
Traffic sign classification
Machine learning
Multi-class classification of traffic-sign images with several algorithms; 93% accuracy on the high-impact classes.
SVM, Naive Bayes, Random Forest, KNN, MLP, PCA
Fall 2021
-
Predicting game success
Machine learning
Poster for Predicting game successModels to predict a new game’s market success, with sentiment analysis of user reviews shown in an interactive visualization.
Linear and Poisson regression, sentiment analysis
-
Face recognition with PCA
Machine learning
Recognizing people from their top eigenfaces, across photos with different expressions.
Python, scikit-learn, PCA
Spring 2021
-
Image compression
Machine learning
Compressing images of several sizes with a K-means implementation written from scratch.
Python, K-means
Spring 2021
-
Spectral clustering
Machine learning
Finding non-convex clusters in a synthetic dataset with spectral clustering.
Python, scikit-learn
Fall 2020
Code for Spectral clustering -
Comparing classifiers
Machine learning
Several scikit-learn classifiers on synthetic data, to show how their decision boundaries differ.
Naive Bayes, trees, AdaBoost, neural nets, SVM, QDA
Fall 2020
-
ISOMAP embeddings
Machine learning
A two-dimensional ISOMAP embedding of face images in different poses, compared with PCA.
Python, ISOMAP
Fall 2020
-
Clustering handwritten digits
Machine learning
Unsupervised pattern recognition on MNIST digits.
Python, K-means
Fall 2020
-
Principal component analysis
Machine learning
PCA on leaf, food-consumption and digit datasets to find the directions that matter.
Python, PCA
Fall 2020
Code for Principal component analysis -
Regularization
Machine learning
Ridge and lasso regression on a housing dataset to curb overfitting.
Python, ridge, lasso
-
NFT analytics dashboard
Data visualization
Analytics and a dashboard over OpenSea trading data and Twitter sentiment.
D3.js, Django, Python, R
Fall 2021
Demo video for NFT analytics dashboard -
COVID-19 data pipeline
Data visualization
An end-to-end streaming pipeline on AWS to analyse trends in COVID-19 data.
Kafka, NiFi, PySpark, HDFS, Tableau
Spring 2021
-
Choropleth map
Data visualization
An interactive map of board-game ratings by region.
D3.js, GeoJSON
Spring 2021
Live demo for Choropleth map -
Co-actor network
Data visualization
A network of actors who have worked together, explored as an interactive graph.
Argo-lite, Python
Fall 2020
Explore graph for Co-actor network -
Force-directed graph
Data visualization
Similarities between board games, laid out with a force simulation.
D3.js
Fall 2020
Live demo for Force-directed graph -
Frequency polygon
Data visualization
An interactive frequency polygon of board-game ratings.
D3.js
Fall 2020
Live demo for Frequency polygon -
D3 line chart
Data visualization
Live demo for D3 line chartBoard-game ratings over time, trying out line-chart features in D3.
D3.js, SVG
-
Tic-tac-toe
Programming
Tic-tac-toe in Python with an AI opponent.
Python
Fall 2020
-
Dice game simulation
Programming
A dice gamble game with Monte Carlo simulation and distribution tests.
Python, Monte Carlo
Fall 2020
-
Fitting distributions
Programming
Routines that fit data to continuous and discrete probability distributions.
Python, goodness of fit