In 2023 I completed a Masters in Electrical and Computer Engineering at University of Washington
As I find time I will try to compile a few interesting examples to high light below.
This was a simple exercise accomplished as part of an Applied Acoustics class at UW. It includes some brief examples of signal processing routines that we discussed in the course.
I am using this space to post selected writeups, code and analysis for some recent projects I’ve been playing around with. Some code can also be found in my github repository.
A simple exercise simulating polling results and looking at the errors. In this notebook I sample from a known binomial distribution as well as a noisy binomial distribution (by simply adding stochastic noise). The ideas in this project were motivated from the book The Signal and the Noise by Nate Silver.
A simple exercise to mutate and model financial data. In this notebook I generate a simple linear models that attempts to predict next day returns from information in the past. As expected, I didn't find much useful signal - e.g. I was unable to usefully predict a future days close from from previous close and volume information. This project was motivated from a chapter in the book Doing Data Science by Cathy O'Neil and Rachel Schutt.
In this notebook I extract and plot Stock ticker data, merging data from different data sources and plotting time series. I compare returns from AAPL and AMZN.
A simple exercise to explore and perform basic inference. Uses the ToothGrowth data set availble in R and investigates the effect of treatment factors on tooth growth response.
As part of an Inference Class this simple simulation demonstrates the Central Limit Theorem. The simulation shows how a sample statistic, the mean \(\bar x\), drawn from an exponential distribution, is itself distributed normally with mean \(E[\bar x] = \mu\) and variance related to the number of samples taken, \(Var[\bar x] = \sigma/ \sqrt{N}\).
A machine learning project based on human sensor data. The goal of this project was to create a supervised learning model for classification of a motion type. The data set comes from this Human Activity Recognition data. The project was completed as part of the Machine Learning class for the Johns Hopkins Data Science series.
A simple exploratory analysis of PM2.5 Emissions data from the EPA National Emissions Inventory. The data was curated and provided by Roger Peng as part of the Exploraroty Data Analysis coursera course. Slicing and dicing this data lead to many simple yet interesting questions and observations.