← All projects

Systems · 2026

GOOBI — Generative Outlier Oriented Benchmark Indexer

Corporate executives, directors, and major shareholders are required by law to report their stock purchases to the SEC within two business days via Form 4 filings. GOOBI (Generative Outlier Oriented Benchmark Indexer) is a financial analytics tool based on the hypothesis that insider transactions will, on average, beat the S&P 500. Goobi uses transaction data from the SEC EDGAR database to perform backend analysis to compute a "quality" score.

Timeline 6 months
Role Lead Software Engineer
Stack Python, SQL
Status In-Progress

Problem

SEC Form 4 filings are public, but the EDGAR system is essentially a raw data dump. An attempt to prove the hypothesis that insider trading typically beat the S&P 500 was made, and proved success; however, manual parsing and analysis of data points proved time-consuming and prone to errors. Continued use of manual labor would prove inefficient and lead to losses in profit.

Solution

In conjunction with my brother, we designed, programmed, and tested a python based analysis tool. This tool parses SEC databases, conducts in-depth analysis, and returns flagged transactions. This system is an automated pipeline that continuously monitors SEC EDGAR for new Form 4 filings, parses and stores the relevant data, and then runs a multi-layered analysis to surface only the purchases most likely to be meaningful. It filters out noise such as grants, awards, and any other purchases which reflect strategic accumulation rather than personal conviction and scores what remains against a set of weighted qualifiers. The output is a ranked, scored list of insider purchases that an investor can act on with genuine context behind each signal.

Impact

The impact of this project is clear, there are several strong paths forward, each with real potential. We’re not limited to a single answer we have multiple good options to choose from. There are a countless number of independant investors most of whom would benefit from a system like this. Currently, talks of monetization systems are in progress, but nothing has been set in stone quite yet.