Does Slightly Mean Somewhat? Measuring Vague Intensity Words in LLM Numeric Actions
Do language models preserve the ordinal meaning of intensity words when those words must produce numeric actions?
Posts by Year
2026
Can Humans Detect AI? Mining Textual Signals of AI-Assisted Writing Under Varying Scrutiny Conditions
Does the threat of being caught change the way people use AI to write? And if it does, can other people tell the difference?
When Treatment Timing Is Messy: Event-Study Design for Staggered Rollouts
You designed an experiment. Leadership signed off. The rollout took on a life of its own. Now what? Introduction All data in this post is simulated for d...
2025
Dimensionality Reduction From Scratch: PCA, ISOMAP, and Eigenfaces
You have thousands of features but only care about the patterns that matter. Here’s how to find them.
Density Estimation and Mixture Models From Scratch
How do you estimate the shape of data when you don’t know what distribution it came from?
Clustering Algorithms From Scratch
You have 10,000 data points. How do you find natural groupings without any labels?
Classification Algorithms From Scratch
Given a data point, which class does it belong to? We build three different classifiers and see when each one shines.
2020
US Hospital Beds: Interactive Map
An interactive map exploring hospital bed capacity across the United States.
Magnus Carlsen: A Career in Data
An interactive exploration of Magnus Carlsen’s chess career through data visualization.
Measuring Urbanicity with Hospital Data: A Distance-Based Approach
How do you measure how “urban” an area is without census data? Here’s one approach: look at where the hospitals are.
Beyond Accuracy: Picking the Right Threshold When Costs Aren’t Equal
Most tutorials stop at accuracy. But what if a false negative costs you 10x more than a false positive?