Additional installation steps and bug for CORD-19 about paperetl

[DavidRivasPhD]

Thank you David. After examining the results more carefully, unlike the above screenshots,
I found plenty on non-NULL values (screenshots below); so the output is correct and there is no such a bug.
Yes, I have read your writings on GitHub, Kaggle, and tds as well as Ankur Mohan’s blogs about your work. Your models are very nice, I have learned a lot from them.

[davidmezzetti]

Hi David,

Unfortunately, NLTK has different models that are installed with different installations. Thank you for sharing that you needed to install an additional model.

Regarding study design, not all documents will have a detected study design (will be 0). Only certain documents will be tagged as COVID-19.

Please take a look at: https://www.kaggle.com/davidmezzetti/cord-19-etl and https://www.kaggle.com/davidmezzetti/cord-19-analysis-with-sentence-embeddings

These notebooks on Kaggle have all the background on how this process works.