ArcGIS Training
In my career research, it seems there are many opportunities if you know how to use ArcGIS. I have done quite a bit of GIS studies during my time at Healthmap, but I never used a licensed software. It’s amazing how some things become so simple with well designed interfaces !
Because these jobs are so interesting, I decided to license myself in GIS through EdX. The learning process has been amazing so far.
Expected Completion : August 2024
Part 1 : GIS Foundations : Status : Complete
Part 2 : 3D GIS
Part 3 : GIS Image Analysis using ArcGIS Pro.
Covid-19 Case Map
This is one of the last projects I worked on before leaving Boston Children’s Hospital during one of the first lock-downs. We worked tirelessly as cases poured in to keep a functional, update, and easy to use data base for public health research. This was a true learning experience for real-time database management, storage optimization, and data pipelines.
My role in this involved installing data cleaning tools as curators typed in new information that could ensure consistency and ease of use. In parallel, I was in charge of the back-end development that leveraged this new dataset to display it on the web map.
If there is one thing that resonates from this project, it is that a good team with open communication makes all the difference.
Though the map has not been kept up to date, you are welcome to visit its remnants.
Connect 4 AI
For this project, I wanted to challenge myself to build an AI using a neural network that could beat me in a classic game of Connect 4.
I began by building a terminal based game engine with a simple AI running off of “if… then…” statements. This was a crucial step to generate game data and start the machine learning process.
With data in hand, I used a convolutional neural network that was essentially treating the game board as an image. Scanning through the different parts of the image, it was eventually able to understand how to win, and started winning consistently against the simple AI. At this point however, it wasn’t enough to beat a human player.
To up its skills, I turned the AI against itself. This widened the variation of moves from its opponent. The bittersweet ending to this story is… it eventually beat me!
To learn more details about the project, I invite you to download the technical report below, or you can go look at the code on GitHub ;)
Feedback mechanisms in the
circumgalactic medium (CGM)
To this day, I consider this one of the coolest things I have ever studied. The basic idea is that as stars die in a galaxy and burst into supernovea, this creates waves of matter and heavy particles that can be shot outside the galaxy (the CGM). This matter is extremely hard to detect due to its low density, and lack of starlight outside the galaxy.
The study focuses on the mechanisms at play that lead to the enrichment of the CGM. Specifically we run grid based simulations to better understand the wind dynamics, and . how star formation rates affect this flow of gas.
Using optimized analytical solutions, I sift through the data to produce representations, and gather insights about our universe. You’re welcome to read the report if you would like to know more =)