Methods
8 articles
47 Scripts to 15: Cleaning a Research Codebase
Research projects accumulate code debt. An AI agent can map dependencies, identify dead code, and reorganize months of accumulated scripts, with counterfactual tests to verify nothing broke.
The Data Quality Problem: How We Went From 49% to 12% Mobility Deserts
We found 49% of California census tracts were mobility deserts. After getting complete transit data, the corrected figure is 12%. Here's what went wrong—and how to avoid it.
6,613 Stores, $147, Zero Lost Data: Robust API Collection
Collecting location data from Google Places API at scale requires handling rate limits, pagination, and failure recovery. A naive script fails in predictable ways.
How to Calculate 2.7 Million Transit Routes for Free
A complete tutorial for calculating multimodal transit travel times at scale using r5py, GTFS, and OpenStreetMap—no expensive APIs required.
400 Labels to 94% Accuracy: Validating Grocery Store Data
Google Places returned thousands of 'grocery stores.' Many weren't. Here's how a classifier separates real grocery stores from gas stations and liquor stores.
7 Copy-Paste Cycles to 1 Command: What Changes with Agent-Based Coding
The difference between chatbot-based coding and agent-based coding is categorical, not incremental. Here's what changes when AI can read your entire codebase.
From Methods Paragraph to Working Pipeline: AI-Assisted Implementation
A well-written methodology section is almost executable code. The gap between describing a procedure and implementing it has narrowed dramatically with agent-based coding tools.
One Context File, Zero Re-Explanations: Teaching AI Your Research Project
Every new coding session with AI starts the same way: re-explaining the project. A single CLAUDE.md file loads automatically every session. Write the context once; the agent reads it every time.