Data Driven Start-Ups

Web scraped text of data-driven start-ups founded between 2010 and 2023. The data was used for data-driven business models analysis, identifying emergent trends and business models transformation over time. The dataset contains the text split into sentences along a reference text (description) and respective embeddings. The foundation model used for the embeddings is: paraphrase-multilingual-MiniLM-L12-v2.

The data collection process not only respected websites' privacy but also adhered to best practices. The scraper tool was configured to read the robots.txt file at the root of each website and proceed only with actions explicitly allowed by the respective site. Additionally, the collection was limited to 50 pages per firm to avoid excessive harvesting.