BoundingPulse

Overview

BoundingPulse grew out of craigSearch, a Python tool that automated and filtered Craigslist searches beyond what the site itself allows. It evolved into a full-stack, AWS-hosted analytics platform that ingests Craigslist posts nationwide for geographic cost insights, with text indexing for advanced search — all backed by a scraping + database pipeline.

Motivation

Craigslist’s own search is limited — there’s no good way to track listings across regions or watch for new posts that match specific criteria over time. I wanted a faster, broader way to monitor public activity across the country. I built the original pipeline back in early 2023, when AI tooling couldn’t carry a full build, so every piece required hands-on problem-solving.

Challenges

• Scraping Craigslist reliably across many cities — Selenium/geckodriver, rate limits, and compliance safeguards.
• Expanding coverage through a city database and moving from a one-off script to a real app backed by a database pipeline.

What I learned

• Designing and deploying a full-stack, AWS-hosted application end to end.
• End-to-end data pipelines and working with a relational database.
• Solving problems with early AI coding tools, back when they needed a lot of hand-holding.

Status & next steps

The core premise — monitoring public activity at scale — still holds, and I like it as a way of surfacing signal from noisy public data. Much of what I learned here carried directly into MarkEverything.