brt-h Goto Github PK

Hello, I'm Robert 👋

I recently finished the IBM Data Science Professional Certification (10-month program), and have been building a portfolio here on github... currently a work in progress!

I'm an experienced account manager and dual major Economics & International Studies from UNC— who is passionate about helping people use data, crypto, and machine learning to create value and solve business problems.

I'm currently looking for a new challenge— so if you've got an idea about me, or know of any good work opportunity that would be beneficial to me, please let me know!

(ie: Technical Account Manager · Product Manager · Account Engineer · Sales Engineer · Software Engineer · Data Analyst · Data Scientist)

Under construction... Python library to for general purpose mouse click accuracy improvement for AI vision models such as GPT-4v, Gemini Pro Vision, Claude 3 and LLaVa. Source code here.

Web application that uses AI to generate a picture book given user inputs, including a description (the main idea of the story) and the total number of pages. Frontend created with Next.js and hosted on Vercel. Backend API written in Python using libraries from LangChain, FastAPI, SSE-Starlette, and Boto3 (AWS)— and deployed to Fly.io using Docker containerization. Live demo, detailed readme, and source code here.

A compilation of 75+ notebooks uploaded as github gists as I worked my way through 10 courses covering python, data science tools, methodology, databases, SQL, visualization, analysis, and machine learning. Since then I have been adding gists for problems I work through on hackerrank, edabit, leetcode, etc.

Enabled a hypothetical Boomerang store to create invoices, track sales team perfromance, and create % change revenue reports. Utilized Excel funtionality such as VLOOKUP, MATCH, helper columns, IFERROR, pivot tables, slicers, group dates, show values as, and more -- with a 50,000 row dataset. Workbook can be viewed here.

Illustrated the drag of inflation on stock market returns compared to Bitcoin. Sourced BTC, SPY, and M2 Money Supply data from Yahoo Finance and St Louis FRED. Utilized three linked data sources with global date range slider. Dashboard can be seen here.

The hypothetical use case for this dashboard is to help someone who is purchasing property in Seattle to rent on AirBnB. The focus of this dashboard is the relationships between AirBnB price and (1) zipcode, (2) # of bedrooms, and (3) week of the year. Dashboard is here.

Predicted between 7 different classes of tree cover found in the Roosevelt National Forest of northern Colorado, using a training dataset consisting of 4 million labled samples with features like elevation, soil type, etc. This location is interesting because of minimal human-caused disturbances, so that existing forest cover types are more a result of ecological processes rather than forest management practices. Utilized Python and Pandas in jupyter notebook to test SGD, LSVC, XGB, CatB, and LGBM classifiers. Kaggle notebook is here.

Enabled a hypothetical competitor to make more informed bids against SpaceX by using 1st stage landing predictions as a proxy for the cost of a launch. Utilized PowerPoint, SQL, Python, JupyterLab, Pandas, Numpy, Dash, Plotly, Folium, BeautifulSoup, Matplotlib, Seaborn, and Scikit-learn. This project served as the capstone for the IBM certification so there are some superfluous slides (14-25) included for peer-grading purposes. Pdf report is here. The code can be found here

Enabled hypothetical stakeholders to explore and manipulate Falcon 9 launch data in an interactive and real-time way. Utilized Python, Pandas, Plotly Express, and Dash. Here is my live dashboard deployed to Heroku. The code can be found here.

Tested four different machine learning models to see which would most accurately predict if a customer would default on their loan or if it would be paid off. Utilized Python, Pandas, Numpy, Matplotlib, and Scikit-learn. Jupyter notebook is here.

Used provided dataset of home sales in King County, USA between May 2014 and May 2015 to practice data analysis. Jupyter notebook is here.

Used yfinance and webscraping to extract stock data then visualize it using Plotly. Utilized Python, Pandas, BeautifulSoup, and Plotly. Jupyter notebook is here.