Legal question in regard to copyrighted reviews.

[jcrodriguez1989]

Dear authors,
thanks for this great script, and for the accompanying publication.
I have a related question. I have web-scraped (through a public, non-documented API) a huge number of individual reviews in the structure of review = (customer, product, rating, review_text). So, these result in short texts (no books, chapters, etc.), what are your thoughts about the legality of transforming individual reviews to bag of words?

thanks,
Dr. Rodriguez, Juan Cruz.

[hawc2]

Hey @jcrodriguez1989, it's legal to transform copyrighted works for non-consumptive research purposes - this is well within fair use. You might like to read Matthew Sag's article, "The New Legal Landscape for Text Mining and Machine Learning.

If you're taking copyrighted reviews and converting them to disaggregated (disordered) bag of words datasets using this script, you can share the bags of words with anyone.

Depending on where you're getting your reviews, they may not even be under copyright restrictions - I would check with the platform/publisher hosting them for their policies on user privacy and what copyright they retain over reviews

[jcrodriguez1989]

Hi @hawc2 ,
thank you very much for your fast reply, and for the provided article.
First, I will try to get the data owner's permission, but I am not even able to get a contact.
I will let you know how it goes.
thanks!