Web scraping developer for property, directory and document data
I build Python scraping and ETL pipelines that pull data from websites, APIs and PDFs, clean it and deliver it where your team already works, whether that is a CRM dashboard or a Google Sheet. At Axiom Realty Solutions, the property-data pipelines I built cut recurring manual work by about 70%.
- ~70% less manual work at Axiom Realty
- Selenium, Apify and Zillow API pipelines
- PDF bank statements to clean CSV
Property-data pipelines at Axiom Realty
Since August 2025 I have worked as a contract Python IT and automation engineer for Axiom Realty Solutions in Chicago. I acquire property data with Selenium, the Zillow API and Apify, choosing per source between a browser, an official API and managed scraping infrastructure. The data flows through ETL into CRM dashboards and Google Sheets, and the dozen external APIs involved sit behind retry, backoff and fault-tolerant fallbacks.
Scrapers for directories and listings
Directory sites hold useful data in inconvenient shapes. My Practo doctor scraper extracts doctor and clinic data at scale, turning listing pages into structured rows. For every source I check the terms first, prefer an official API when one exists and keep request rates polite. Then I pick the lightest tool that works reliably for that source:
- Selenium for JavaScript-heavy pages
- Apify for managed, repeatable runs
- Official APIs, such as the Zillow API, when available
- Retries and logging so broken selectors surface quickly
Getting tables out of PDFs
A lot of business data arrives as documents, not web pages. PDF2CSV-AI, a tool I built, converts PDF bank statements into clean CSV using deep-learning table extraction, so rows and columns survive layouts that break simple text parsing. For screenshots and photos I have built on-device OCR, including Hindi Devanagari, for Life Search AI.
Transform, load and report
Extraction is half the job. I clean and normalise data with Pandas, remove duplicates and load it where it will be used: Google Sheets and Workspace APIs for reporting, CRM dashboards for the people who act on it, or MongoDB and PostgreSQL when it needs a real database. Runs can be scheduled on a Linux VPS or orchestrated through n8n.
Rates from $20/hour
Hourly for open-ended work, fixed price for defined projects. The first call is free and there's no obligation.
Frequently asked questions
Can you scrape any website?
Most sites can technically be scraped, but I check each site’s terms first and use an official API or licensed data source where one exists, as with the Zillow API at Axiom. When scraping is appropriate I use Selenium or Apify and keep request rates reasonable.
What happens when the source website changes?
Pipelines are built with retries, fallbacks and logging, so a changed page shows up as a clear failure rather than silently missing rows. Fixing a broken selector is then a contained repair instead of a data-quality mystery.
Can you extract data from PDFs and scanned documents?
Yes. PDF2CSV-AI turns PDF bank statements into clean CSV with deep-learning table extraction, and I have built on-device OCR, including Hindi Devanagari, for Life Search AI. Tell me about your documents on a free call and I will say what is realistic.
How much does a web scraping project cost?
From $20/hour, or a fixed price once the sources, fields and delivery format are defined. Ongoing pipelines that need monitoring and repairs can continue on the same hourly rate, billed only for the hours the work actually takes.