SEC EDGAR Data Pipeline & Earnings Analytics

A 7-script Python pipeline that pulls XBRL financials, 8-K earnings dates, and daily prices for 18 software companies plus the S&P 500 (SPY) from the SEC EDGAR API and Yahoo Finance into an indexed SQLite schema.

Problem

Comparing software companies meant manual 10-K pulls, and XBRL data is inconsistent: tags differ between filers, restated quarters repeat, and some multiples are meaningless. The comps had to reconcile four-quarter LTM figures to annual 10-K totals.

Impact

Revenue growth explained valuation multiples about 3.8x more than margin did. Earnings-day moves (avg. 8.1% across 469 releases) were not predictable out-of-sample from pre-announcement signals.