We have the transcript for millions of text files (range of a few terabytes). We want to make a statistic of what our users answered to a set of questions. We do not necessary look for natural language processing because we are looking for keywords after the question has been answered (regex).
Example of a question: Are you interested in this monthly subscription? Answer: No, not at this time.
We prefer to pay extra on Cloud (more processing power).
Can Node.js do the job?
What Node.js package do you recommend?
As an server, is it enough a1.4xlarge 16 (vCPUs) 32(GiB) from AWS? We only need to process the files when we desire.