Problem Statement: Bitcoin and Ethereum are two of the largest cryptocurrencies. These are digital money that runs independently of any central authority or government monitoring. Peer-to-peer software and cryptography are used instead. These were founded with the intention of allowing people to send money over the internet.
In this project, we're attempting to determine whether global news i.e., tweets concerning Cryptocurrency-related issues has an impact on the price of a Bitcoin and Ethereum. Our project aims to forecast the change in the dollar value, i.e., whether the price will rise or decline in response to the world news.
Implementation:
Data Extraction: We have imported the Bitcoin and Ethereum data from OpenBlender into Panda Data frames.
Exploratory Data Analysis: Post extraction we have calculated several insights out of the original datasets using basic logarithmic and plotting functions and computed a target variable.
Data Blending: Further we have blended our cleaned data with the dynamic cryptocurrency data available on OpenBlender using the time stamps and the tweets that were made in the news about the cryptocurrencies on that particular day.
Polarity Evaluation: We used this blended data to calculate a polarity score of a tweet for a particular timestamp, categorized it as positive negative or neutral and established a relationship between the variation of price and the polarity score on that date and analyzed those visualizations.
LSTM Model Building: Next, we established an LSTM model for each cryptocurrency that is Bitcoin and Ethereum. Splitted the data into training and testing. Fitted the model and calculated the performance measures.
Cryptocurrency Prediction: Lastly, we used the established model to predict the changing rates of Bitcoin and Ethereum cryptocurrencies over the next 60 days or for any time frame as per our desire.