Predictive Models

Air BnB Price Predictor

Worked on a team to create a neural network that predicts the price of an Air BnB.

I worked with two other data scientists to create a model that predicts price of an Air BnB. I believe that model creation is not something that can easily be done by multiple people, and we were finding that there was not enough work for us all to do something meaningful for the project. Instead we made it a bit of friendly competition, and we all created models and used the model that had the lowest MAE (mean average error). Unfortunatly my model did not win the competition, but you can find it below if you are curious at looking at it.

The repo for the project can be found here on github.

The repo for my personal model can be found here on github.

Adult Diabetes Predictor

Using an XGBoosted regressor, trained a model to predict the % of adults with diabetes.

A toy app of the model can be found here on Heroku.

A blog post I wrote about the work can be found here on Medium.

The repo can be found here on github

Data Exploration

C4ADS Nuclear Fuel Cycle

I worked on a project for C4ADS. The goal of this project was to create a model that could predict whether or not a company had the capability to create precision materials that could be used in the creation of nuclear fuels or weaponry. I made a longer post about his that can be found here on my website

Drug abuse data storytelling project

Looked through the CDC’s Multiple Cause of Death (Detailed Mortality) and found some disturbing statistics.

A blog post can be found here on medium.

The repo can be found here on github.

App Creation

Med-Cabinet Strain Recommender

Worked on a team to create a python backend that took a bunch text, used NLP to create a prediction, and returned Medical Marijuana strain recommendations.

I worked on creating an SQlite database with all the strain information and taking the prediction, querying the database, and returning the strain information in a JSON object for the front end.

The DS repo for this project can be found here on github