Two things happened to me around age 11: I discovered hacking and began teaching myself how to program. Although my programming began with writing video games on my graphing calculator, eventually I moved on to other languages and entered college for computer engineering. Somewhere, I took a wrong turn and ended up a lawyer preparing Japanese companies for antitrust litigation. I am ecstatic about having returned to my first love of data science. In my spare time, I race track bikes and hope to win the national championship next year.
Data Scientist @ Currently working on using Syrian Civil War data to predict war crimes.
Automatically identified which countries a government document is about using named-entity recognition as labels for an SVM classifier.
Wrote a twitter-bot that finds the sentence in a text corpus most similar to a user’s tweets using a NoSQL database and doc2vec.
Created an interactive map and chart in D3 based on Ashley Madison user data. http://gifsandgraphs.com/Dear-Ashley-Madison/ From July 2015 to Present (4 months) Data Scientist @ Used MTA data to optimize deployment of a non-profit’s street team.
Scraped website data and used text mining in order to show the effect of violent media consumption on violent crime using linear regression. Created a script to help BeautifulSoup spider entire websites offline, negating the need to worry about an IP address being blocked for a DoS attack.
Calculated the contributing factors for heart disease using supervised machine learning classification. Wrote a program to brute-force analyze multiple models across differing hyper-parameters. Created a D3 visualization illustrating the accuracy of various models and features for predicting heart disease. http://gifsandgraphs.com/Heart-Disease-D3/
Predicted the passage of federal legislation using natural language processing, logistic regression and clustering. Used unsupervised machine learning to train custom sentiment analysis.
Used text processing and dimensionality reduction with word2vec to predict restaurant health code violations with Yelp data. Ensembled multiple models to increase accuracy. Model tested and performed well in a data science competition. http://gifsandgraphs.com/Winning-Kaggle/ From January 2015 to June 2015 (6 months) Japanese Trial Preparation Attorney @ Worked for various law firms, including Skadden Arps, Clifford Chance, Ropes & Gray, and Hogan Lovells. Worked on a variety of cases preparing Japanese written materials for antitrust litigation with the Department of Justice. Cases covered a broad range from medical patents to video game code. From 2008 to 2014 (6 years) Technology Coordinator @ Designed and wrote the interactive organization website. Created internal web-based applications in PHP, including a content management system for the research organization staff. From 2003 to 2004 (1 year) Computer Consultant @ Designed and wrote the company website. Troubleshooted and rewrote company logistics software using an unknown programming language. From 1999 to 2000 (1 year)
Data Science @ Metis From 2015 to 2015 JD, Japanese Law @ Temple University - James E. Beasley School of Law From 2006 to 2007 JD @ Syracuse University College of Law From 2004 to 2007 BA, Compute Engineering, International Affairs, International Economics, Japanese @ The George Washington University From 1999 to 2003 Japanese Studies @ Kansai University From 2001 to 2002 Aaron Mangum is skilled in: Data Science, Python, Machine Learning, Data Analysis, Amazon Web Services..., Japanese, Mergers, Legal Research, International Law, International Relations, Intellectual Property, Django, HTML 5, Git, Github, CSS, jQuery, Objective-C, iOS development, PHP, MySQL, Web Scraping, Mergers & Acquisitions, Statistics, SQL, Analysis, NumPy, SciPy, Hive, Hadoop, Apache Spark, Management, Project Management, Social Media, C