I have written a program that scrapes data from the web and I have in possession about 5k sentences which I want to analyze.

Part 1: I am just starting out in data science and wanted to know if there is such a thing as a library that will read through text and automatically classify words or sentences as being positive/negative/neutral. Is there such a thing that exists, perhaps in R or Python?

Part 2: From what I can tell, there isn't such a library, and I would have to do that myself by hand. It appears that the best way to analyze sentiment is to:

  • a) put the words in a corpus
  • b) do things like word stemming and clean up the text
  • c) download a dictionary of predetermined sentiment
  • d) run a program that vectorizes the text
  • e) compare the text from my corpus to the dictionary
  • f) ..... unsure what after this step.

Are these steps accurate, or am I totally off-base?


2 Answers 2


Here is a tutorial, that should help you get started.


Part 1: There are pretrained state of the art models trained on their corpus which could be used for sentiment analysis.

Part 2: You may always train your data on their model rather than creating your own model.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.