Svetlana K.

588,000+ business records. 59,000+ restaurant records. 45,000+ location records across 34 countries. I build custom Python scrapers and data pipelines for websites where extensions, no-code tools, and simple scripts are not enough. My projects cover directories, e-commerce, healthcare, public records, location data, restaurant listings, news archives, and document extraction. I manage each project from collection through final delivery. SELECTED RESULTS • Processed 134,000+ restaurant listing URLs for a national dataset • Consolidated 43,000+ healthcare professionals and organizations from 26 sources across four countries • Processed 116,000+ company-record rows using website extraction, PDF processing, and OCR • Collected 35,000+ e-commerce products with prices, descriptions, images, and source URLs • Prepared products, variants, images, pricing, descriptions, and metafields for Shopify import SERVICES • Business directories and company databases • Products, prices, SKUs, variants, and inventory • Location, healthcare, restaurant, and hospitality data • Real estate, court, probate, and public records • News and publication archives • Search forms, filters, pagination, and profile pages • XML sitemaps and large URL collections • Public APIs and client-authorized website endpoints • Authorized login-based workflows BROWSER AUTOMATION I work with websites that require JavaScript, infinite scrolling, browser sessions, cookies, interactive forms, or multi-step navigation. My stack includes Playwright, Camoufox, Selenium, Selenium Wire, Scrapy, Requests, aiohttp, curl_cffi, and BeautifulSoup. I use direct requests when possible and browser automation when required. PDF AND OCR I download documents, extract text from PDFs, process scanned pages, apply OCR, locate fields with validation rules and regular expressions, and combine document data with website records. My document stack includes Tesseract, OpenCV, pdf2image, Pillow, and NumPy. DATA QUALITY • Deduplication • Address and phone normalization • Missing-record detection and source validation • Country and location filtering • Multi-source consolidation and consistent schemas • Recovery passes for failed records • Translation with local caching • Checkpoints and periodic saving DELIVERY • Excel, CSV, and JSON • Shopify-compatible CSV • Files prepared for database import For long collections, I use asynchronous requests, multiprocessing, retries, caching, checkpoints, crash recovery, and chunk-based processing. Temporary errors do not force a full restart. I work with publicly available data and client-authorized sources, using access methods appropriate for each project. HOW I WORK I inspect the source and choose the most reliable collection method. I confirm the fields, volume, output format, and whether you need a one-time dataset or reusable code. For larger projects, I deliver a small initial batch after the contract begins so you can verify the structure and data quality before full collection. I build the scraper and check the output against the agreed requirements. I deliver reusable Python code and setup instructions when included in the project scope. GOOD FIT • Large datasets • JavaScript websites or interactive forms • Scrapers that stop or miss records • Extraction combined with data cleanup • Reusable Python code Send me the target website. I will review the source and recommend a practical approach before development begins. Feedback from one of my client: 📍"Svetlana is AMAZING! I have a fair amount of data background and I am very picky about data quality. She delivers excellent quality data in a timely fashion. I spot checked the data that she delivered and it all checked out well. In addition, when we hit an issue, she offered proactive and creative solutions. I highly recommend her. She is your data wizard, don’t go any further!" Mary, US/ Jul 2025 I am always available to discuss all the details of your project!