Aaron Abrahamson

Data Scientist

Hi there! I am a Seattle native, a graduate of the Data Science program at the Flatiron School, and I have over 4 years experience as an analyst working in the financial services industry.

This page lists some of the various projects I have worked on, as well as some articles I have written.

If you'd like to get in contact with me, please send me an email: aabrahamson3 at gmail


Projects

Beer:30

A brewery recommender for travelers

View Project Website
View Project Github

Beer30 is powered by a deep neural network (Doc2Vec) and is trained on over 2 million text reviews spanning from 1998-2017. The model learned from the taste and flavor notes provided by over 20,000 users on 22,514 unique beers. It takes in either a keyword or specific beer, and it searches for other beers most similar to the provided input. Then it filters based upon the location provided by the user, and returns breweries that make beer that closely match the keyword/beer.

Council Data Project

Participated in the 2020 St. Hack-trick's Day hackathon

View Project Github

The Council Data Project is aimed at making city council meetings more transparent and discoverable for citizens and journalists.

I performed preliminary EDA on text transcript data from Seattle City Council meetings, and then created a pipeline to evaluate and visualize sentiment of meetings over time with NLTK's VADER.

Predicting Faulty Water Wells in Tanzania

Using Machine Learning to detect when Tanzanian water wells are in need of repair

View Project Github

Water is an incredibly volatile resource in Tanzania - nearly half of the population does not have access to clean drinking water. We analyzed data provided by the Tanzanian Ministry of Water on 60,000 water wells (and their operational status) and used it to build models with the aim of predicting when they are in need of repair. We created a data pipeline to extract and engineer complex features to improve model’s predictive accuracy. We selected a random forest model with an accuracy of 82% (F1 scores 79-84%)

Reconnecting to Opportunity

Estimated and revised the number of Opportunity Youth in South King County

View Project Github

Opportunity Youth are defined as young people aged between 16 and 24 who are not participating in the labor market or are not enrolled in school. Approximately one in every nine people in this age group is considered an Opportunity Youth, and their chance of earning an associate's degree (or higher) is severely diminished.

Current estimates of Opportunity Youth in King County are based upon 2014 census data, but there are now more up to date figures. This project sought to re-estimate the number of Opportunity Youth in King County based upon this new data, to help social programs aimed at aiding these individuals to succeed.

Helping First Time Home Buyers

Modeling King County Housing prices to help Millennial first time home buyers

View Project Github

Created a linear regression model to test several hypotheses about housing prices, with the aim of helping Millennial first time home buyers.


Articles

Detecting Fake News With Deep Learning

A Simple LSTM Implementation With Keras

View Article

I wanted to try out both a supervised NLP problem and a LSTM model. So I did both!

Analyzing Sentiment of City Council Meetings

Can sentiment indicate noteworthy portions of a meeting?

View Article

Walking through the work I did during St. Hack-trick's Day 2020.

Creating a Brewery Recommender with Doc2Vec

How I trained and evaluated my capstone project, Beer:30

View Article

In this article I go through how I created a content based recommender with Doc2Vec, and I show some cool visualizations!

Weaponized Natural Language Processing

Automated end-to-end spearphishing on twitter

View Article

In this article I describe and summarize a project created by a team of security researchers. They developed a way to automate spear phishing campaign's on Twitter. It is an incredibly interesting application of natural language processing.

What is Docker, and why is it useful for data science?

An overview of the container platform, Docker

View Article

Docker is a great way to encapsulate data science projects. In this article I give a brief introduction on what it is!

How To Run Docker Images and Make Your Own!

A tutorial on how to run docker images and make your own

View Article

This is a technical tutorial on how to run a Docker image, make your own Dockerfile and turn it into a Docker Image!