Name: Bitcoin 17.7 Million Tweets And Price
Creator: Kaggle
License: https://creativecommons.org/publicdomain/zero/1.0/

About this Dataset

Bitcoin 17.7 Million Tweets And Price

Small description

This dataset contains the average sentiment of all tweets about bitcoin from 01/08/2017 until 21/01/2019. It also contains the finantial data of bitcoin for that same period.

How did I gathered the tweets?

To collect all tweets I used this github. Some days have missing data but I think it is minimal. I collected over 17.7 million tweets

How did I do sentiment analysis?

I used the library VaderSentiment. I added about 30 expresions and words to the dictionary. To score the expresions I used the same methodology as the authors described in their paper.

What do I want to accomplish?

I would like to create a predictive model for bitcoin's price using only twitter's data. Check the thesis paper here

Tables

Data To Hourervals No Filter

@kaggle.jaimebadiola_bitcoin_tweets_and_price.data_to_hourervals_no_filter

1.04 MB
12,936 rows
14 columns

CREATE TABLE data_to_hourervals_no_filter (
  "date" VARCHAR,
  "compound_score" DOUBLE,
  "n" DOUBLE,
  "count_negatives" DOUBLE,
  "count_positives" DOUBLE,
  "count_neutrals" DOUBLE,
  "sent_negatives" DOUBLE,
  "sent_positives" DOUBLE,
  "open" DOUBLE,
  "high" DOUBLE,
  "low" DOUBLE,
  "close" DOUBLE,
  "volume_btc" DOUBLE  -- Volume (BTC),
  "volume_currency" DOUBLE  -- Volume (Currency)
);

Df Final

@kaggle.jaimebadiola_bitcoin_tweets_and_price.df_final

1.1 MB
12,936 rows
16 columns

CREATE TABLE df_final (
  "date" TIMESTAMP,
  "compound_score" DOUBLE,
  "total_volume_of_tweets" DOUBLE,
  "count_negatives" DOUBLE,
  "count_positives" DOUBLE,
  "count_neutrals" DOUBLE,
  "sent_negatives" DOUBLE,
  "sent_positives" DOUBLE,
  "count_news" DOUBLE,
  "count_bots" DOUBLE,
  "open" DOUBLE,
  "high" DOUBLE,
  "low" DOUBLE,
  "close" DOUBLE,
  "volume_btc" DOUBLE  -- Volume (BTC),
  "volume_currency" DOUBLE  -- Volume (Currency)
);