Context The data was scraped with the aim of training an algorithm which can code medical diagnoses without human intervention gaining the expert knowledge from the webpages scraped from the website. Content 1. **level_4_links** are the links for the pages corresponding to a given ICD code in the website [ICD10Data.com][1] 2. **bs4_l4_dump.tar.gz** has the **92389 scraped pages** as per the existing codes at the time of scraping. 3. The code for the scraping can be found at [https://github.com/shams-sam/ICD10Data.com][2] [1]: http://icd10data.com [2]: https://github.com/shams-sam/ICD10Data.com Acknowledgements We wouldn't be here without the help of others. The contribution of the website in curating and maintaining the website without any broken links is truly commendable. Inspiration We have already come up with some of the best results to code the diagnoses strings using LSTMs and word embeddings. Please go ahead and give it a try.