Carson Weeks: Projects
Hello, I'm Carson Weeks, an Economics major with a passion for blending the realms of economics, data science, and markets. My goal is to harmonize these disciplines to elevate my hobbies and achieve my ambitions, forging innovative pathways in the process.
Portfolio
Market Microstructure Analysis
This project gave me a hands-on look at what actually happens inside financial markets, beyond the price charts most people see. I worked with order book data from a crypto exchange — essentially a record of every bid and ask — and tried to piece together how liquidity, spreads, and trading depth play into execution costs. The raw data was messy and extremely high-frequency, which meant I spent a lot of time cleaning and organizing it before I could even start analyzing.
Once it was ready, I dove into questions like: how much does latency matter in trade execution? How do small shifts in liquidity affect price impact? What stood out to me was how sensitive outcomes can be to micro-level frictions — things that most traders never see but that drive the efficiency of a market. Working through this made the abstract idea of “market structure” feel much more tangible.
Trading Strategy Backtesting Engine
This project was my attempt to bring trading ideas into the real world without putting any money on the line. I built a backtesting tool that could run momentum and mean-reversion strategies against past market data, and I wanted it to feel more like a sandbox than a rigid program. You could tweak parameters, rerun the tests, and quickly see how things like Sharpe ratio, drawdowns, or overall returns shifted. It was rewarding to watch strategies come alive in that way.
What surprised me was that the real challenge wasn’t writing the strategies — it was making sure the results weren’t giving me a false sense of success. I had to factor in things like transaction costs and slippage, which don’t seem huge at first but can completely change whether a strategy is viable. By the end, I had something that felt honest and realistic, not just numbers on a screen. More than anything, it showed me how easy it is to be fooled by clean data, and how much discipline it takes to tell the difference between signal and noise.
Non-extreme Weather & Economic Activity
For this project, I wanted to understand how something as unpredictable as the weather can ripple through an economy as complex as New York City’s. I pulled together detailed records on temperature, precipitation, et cetera, and matched them with taxable sales across different industries. The challenge was figuring out how to separate “noise” from actual economic impact — for instance, whether a rainy weekend really lowers retail sales, or if the effect disappears once you account for other factors.
To get there, I used panel regressions with fixed effects and instrumental variables, tools that help cut through the clutter and isolate causal effects. What I found was that weather shocks do meaningfully shift consumer behavior in the short term, but the degree of impact varies by sector. The process taught me a lot about the intersection of econometrics and real-world data: the numbers don’t just tell a story on their own — you have to work carefully to pull it out.
