Geospatial real estate database, predictive model, and dashboard
Zillow? MLS? This platform integrates local property data with geomapping, providing an innovative real estate analytics dashboard. I personally used this to find and purchase my first house in a highly-competitive market back in 2019.
​
Utilizing Python and Selenium, I created a "crawler" program to extract real-time data from county auditor systems and compile a semi-structured database.
After that, I used OpenStreetMaps for geocoding and scikit-learn to develop a predictive model. This model could identify off-market properties likely to sell, based on historical patterns. To make these results more accessible, I developed a Shiny app for data visualization and contacted properties that piqued my interest.
​
The results: I purchased an excellent first home at 60% the bank appraised value within 4 months of starting the project.
Global freight movement analytics platform
Collaboration with a successful freight broker (my brother-in-law) identified a opportunity to leverage global data sources in a novel manner.
​
Brokering freight is intensely competitive, and success is largely dependent on one's ability to identify new business opportunities.
To find these business opportunities, I thought we might use some data.
I brokered a deal with an international data provider for global port manifest data of goods entering and exiting the United States.
Using Natural Language processing, record linkage, and Python, I processed this data into a structured relational database of companies, contacts, shippers, goods, ports, and lanes.
The processed data was still too large to interact with using typical business tools, so I created a platform using shiny apps and Google Cloud Platform for an easy-to-use interface that enabled my brother-in-law and his team to find and analyze new business opportunities.
​
This has resulted in a tremendous increase in the identification of qualified business leads, many of which have been converted to lucrative customers.
Bible GPT
This Flutter app is an experiment with the application of AI and matters of faith.
BibleGPT uses an OpenAI gtp-3.5-turbo model to provide a conversational chat experience, leveraging a custom knowledge vector database of scripture, church doctors, and faith resources served through a flask api.
​
For asking the BIG questions, like, "is it wrong to use AI technology to interpret scripture and theology?"
"Historic International Movie Data Base", or, "HIMDB" is a fun project that I put together to indulge myself as a cinema and history fan.
​
Using the open IMDB database, I created a collection of movies with historic context and used OpenAI's api to enrich the data with the film's plot location, historic events, characters, and time era. I used this geocoded this data and created a shiny app to easily explore cinema using point-in-antiquity and location navigation of the film's plot. I also created a basic api for linking the amazon prime video titles using a SERP api and url parsing.