Find More About Me at James' Portfolio Site
I am a team-oriented data scientist who creates relatable stories using machine learning models and visualizations via python to develop insights and help solve problems with data.
[
- Querying: PostgreSQL / SQLite
- Coding: Python
- Analysis: Excel / Pandas
- Visualization: Matplotlib / Tableau
- Machine Learning: Sklearn / Tensorflow 2.0
- Big Data: Spark / Big Query
- Workflow: Bash / Git
- GeoSpatial Kepler.gl, Geopandas
- NLP SpaCy, NLTK
- Wine Varietal Predictor: Modeled a 21-class predictor using a convolutional neural net that was able to predict wine varietals 3 times better than the baseline.
- Auto Accident Severity Predictor: Deployed a webapp that predicted auto accident severity as part of a team working on a data set with over 4.2 million observations.
- Social Media Classifier: Extracted 6000+ social media posts from two different Reddit threads and processed them through a Natural Language Processing workflow that used a Voting Classifier model to classify them with a balanced accuracy of over 90% versus a baseline of 27.5%.
- Real Estate Price Predictor: Analyzed a housing dataset with over 82 features to identify relationships and build a predictor model that could determine a housing price within less than 10% of its actual value.
- STEM-Readiness Analysis: Analyzed National and State of California ACT data to find relationships between size and type of school district and the influence on ACT science and math scores.
- Distribution Expansion Analysis: Analyzed demographic information of 6 potential regions to determine the best location for a regional distribution hub.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent