Data Pipeline · Case Study · Pelumi Oladokun

Data Pipeline

X Profile Scanning Pipeline

A vetting workflow that ran for two days unattended and returned 20,509 X profiles with verification status, follower counts, and region data filled in.

X Profile Scanning Pipeline project visual

The brief

What needed to be solved.

A Web3 marketing team had more than 20,000 X usernames and needed three things for each one: verification status, follower count, and region. Manual checking would have taken weeks.

The hard part was not opening one profile. It was keeping a long run alive, saving progress, and ending with a spreadsheet the team could use right away.

The constraint

What made it interesting.

I used Playwright instead of the paid X API because this was a one-time job and accuracy mattered more than speed.

The pipeline saved progress every 25 profiles, added random delays, retried failures, and skipped media downloads so the run could stay stable on a normal laptop.

The build

What was assembled.

Checkpoint files so the run could restart without losing progress.

Random delays and timed pauses to reduce rate-limit risk.

Retry logic for broken pages and temporary failures.

Spreadsheet output with filtered views for campaign use.

The result

What changed after it ran.

20,509 profiles scanned in one unattended run.

395 verified accounts found and 10,192 dead accounts flagged.

The client saved outreach time because the scan cleaned the list while enriching it.

Stack