From fund reports to validated holdings

I built FundFlow to automate a recurring operations task: collect a fund’s holdings report, check its contents, and make the results available for review.

Using real iShares reports, the pipeline preserves original CSV files in an AWS S3 bucket, validates their contents, and stores accepted holdings in MongoDB Atlas with a link to their source. Explore the reports and processing history below.

Built with Python · Playwright · AWS S3 · MongoDB Atlas · FastAPI · pytest · Vercel

Fund overview

The latest processing attempt for each fund.

Loading reports…

Fund overview
Fund	Report date	Accepted	Rejected	Latest outcome	Reports

Processing history

Up to 20 recent attempts, newest first. Times are shown in your local timezone. Accepted counts describe validation; inserted and unchanged describe acknowledged writes.

Processing attempts
Started	Outcome	Accepted	Rejected	Inserted	Unchanged	Run ID

Holdings

Latest successfully processed report. This may be older than the latest attempt shown above.

Report holdings
Name	Asset class	Market value	Currency	Weight (%)