The real problem
Building a usable lead list required repeating searches across keyword and location combinations, then cleaning inconsistent records before they could be used.
Python · Playwright · Google Sheets
Automated Lead Pipeline
I built the original pipeline to turn keyword and location combinations into a structured, resumable lead dataset. This public recreation uses fictional businesses and a simulated directory so the workflow can be explored safely.
2,700 planned searches
Cache and resume
Automated sheet output
Public-safe demo. It does not contact Google Maps, expose credentials, or use real lead data.
Python automation
I built it to turn search lists into usable lead data.
This recording shows my Python and Playwright workflow processing a fictional search queue, cleaning the results, and writing them into a structured table as it runs.
Behind the project
A real data pipeline, recreated with safe inputs.
What I built
I built a resumable Python and Playwright pipeline that processes a search queue, collects records, normalizes fields, removes duplicates, and writes structured results.
Why this version is different
This version uses fictional businesses and a simulated search process. It shows the queue and output behavior without publishing source details, credentials, API keys, or private lead data.
Technical focus
Queue management, browser automation, normalization, deduplication, checkpointing, resumability, and structured output.
Input
Search queue
Edit the fictional keyword and location combinations before running the workflow.
| Status | Keyword | Location | Remove |
|---|
4 searches ready
0 cached
Automation
Run progress
Waiting to run
Directory Search
No active query
⌕
Results appear here while each search is processed.
Output
Lead results
A sheet-style preview of the normalized and deduplicated output.
| Business | Category | Location | Phone | Rating | Source query |
|---|---|---|---|---|---|
| Run the pipeline to generate fictional lead rows. | |||||