Blog
Everything here is something I found interesting enough to want to write down. Sometimes it turned out to be useful for work; that was never the reason it got posted. The bar is whether I think someone else might find it interesting too.
The topics range widely because where I live they have to, and because the underlying math tends to be the same across them anyway. A lot of what I end up working on are things the industry overlooks, namely problems that are real but not profitable enough for someone else to build the tools for. I'd rather those tools exist than not, and I like building for the people who actually use them.
Most of these posts sit somewhere between a research paper and production code. Papers tend to stop once the result is established; production code tends to start once the choice has already been made. I'm usually more interested in the space between, like what a technique actually implies once you take it seriously, where else it might apply, which assumptions matter and which ones were incidental. That's the part I find worth writing down, partly because it's the part that's hardest to find written down anywhere else.
High Level Analysis of Race Results
Providing insights for spectators of races at scale
Tags
Predicting Attributes/Demographics
Using probability to estimate demographics while managing uncertainty
Tags
An Introduction to Optimization: Combinatorial Optimization
Getting Started with solving combinatorial optimization problems
Tags
Web Analytics for Race Result Websites
Identifying and recording data points about visitors on race result websites for track & field and road races to identify behaviors, needs, and desireable features of users
An Introduction to Optimization: Linear Programming
Optimizing continuous variables based on constraints using Google's Operations Research library for Python, ORTools.
Tags
6 simple applications of Machine Learning for running
Writing machine learning models in Python's scikit-learn for gathering insights into sports
Tags
Structuring Donor databases as networks
Determining the dynamics of your donor network
Tags
Trajectory Analysis
How to extract information from detail trajectory data, and compare trajectories.
Tags
Pack Membership and Roles in Race Competitions
Using pack membership, roles, behavior, and changes to make explain race outcomes in running, cycling, speed skating, cross country skiing, etc
Tags
Network Science - Node Strength, Centrality and Communicability
An introduction to Hirsch index-based metrics for measuring node importance in weighted graphs based on degrees and node strengths, and a centrality for measuring communication ability of a node in a network.
Tags
Compositional Data Analysis
Ratios and logratios are the standard metrics for working with compositional data and are the only metrics that allow for statistical analysis of compositions.
Tags
Data-driven Track & Field/Cross Country
Metrics for more mathematical approach to evaluating and quantifying track & field performances, the competition, and areas for improvement.
Tags
Data Simplification & Summarization
Simplifying polylines and trajectories using the Ramer-Douglas-Puecker algorithm, and summarizing graphs/networks to speed up computations using grouping-based and compression-based algorithms.
Tags
Balancing LLM Prompt Analytics and User Privacy
Learn how to implement effective prompt analytics for large language models while safeguarding user privacy. Discover key strategies, best practices, and ethical considerations in this guide.
Tags