Explanation of “BITCOIN TWITTER Sentiment Analysis” Automatic posts

Dear Readers/Followers/Investors.

With this short article I hope to be able to give some clarifications about the “BITCOIN TWITTER Sentiment Analysis” automatic posts that we publish every day both here, and on steemit.com.

I use a lot twitter, and in particular crypto twitter, actually I use twitter just to get cryptocurrency related posts, and infos, and nothing more.


Because of my programming background, and my recent work with artificial intelligence, I started to be interested in building a classifier (a database to train A.I.) to perform a sentiment analysis on the daily posts about crypto, and BITCOIN in particular.

I thought this could be interesting to get some “hidden” information, otherwise invisible at first glance.

In order to do so I proceeded as follow:

  • Code a program to extract all the posts (or the majority of them) with the keyword BITCOIN and save them on a local database.
  • Once a certain number of entries have been reached, manually categorize them into SPAM, and GOOD tweets.
  • Code an AI predictor trained on the previously manually created classifier.
  • Once the predictor reached an acceptable level of accuracy, I upgraded the AI code to include the paralleldots.com sentiment analysis API.
  • After that, every new tweet posted, get saved, and analyzed by my codes, and classified as SPAM or GOOD, and the GOOD ones get analyzed and divided by sentiment, than stored in my databse.

With all of these data properly classified, I’ve been able to code another program to create 3 different plots, extract the most used keywords, and from these information create the unique content, updated every 6 hours, that gives to you, we think, the useful information about the current sentiments around cryptocurrencies.

Information that are useful, in my opinion, to help a careful observer to create a better vision around what is going on, and what potentially could happen, in the crypto community.

Thanks for reading

This post is published on steemit as well 

Please follow and like us: