1. Generate a list of company names from AngelList and Crunchbase.
2A. Automate Google searches for "[company name] + [valuation]." Scrape search results.
2B. Alternatively, use Google Custom Websearch API and search for "[company name] + [valuation]."
https://developers.google.com/custom-search/json-api/v1/over...
3. Parse search results by writing a regular expression that matches common patterns of how a company valuation might be phrased.
Examples: "at $10B valuation," "valued at $10B", "valuing the company at 10 billion"
4. Sanitize data manually, or outsource to Mturk to get people to validate results.
Thanks anyhow.