Bitcoin prices are partially set at the watercooler. Since most cryptocurrencies have yet to reach meaningful adoption, market movements depend on the whims of the investing public. Questions like “do most people expect Bitcoin to crash?” have more weight for short-term movements than the technical specifications of PoW mining or the Lightning Network.
Every day Twitter users post thousands of messages about Bitcoin and cryptocurrencies, and these tweets can collectively represent market sentiment. Although it’s difficult to quantify, these messages can provide a bird’s eye view of the investing crowd – assuming you could filter out the spam, that is.
That’s where sentiment analysis comes in. Sentiment analysis mines text from the web and extracts subjective information about the words being used, like ‘HODL’ or ‘Moon’ or even ‘Rekt.’ It uses a sliding scale to determine whether the tweets of other media are positive, neutral, or negative, while simultaneously using logic to ignore spam accounts.
Take the following two sentences for example. We can use the program Parallel Dots to determine their sentiment:
- Bitcoin is going to the moon! I love it!
2. Bitcoin is worthless. It will never scale.
There are approximately 22,000 tweets about Bitcoin per day, according to Bitinfocharts. Sentiment analysis tools take all of those tweets, compile their sentiment data, and spit out a number that represents how people feel about the largest cryptocurrency.
What does this mean for the market?
Circle Research analyzed data from The Tie, an analytics company focused on predicting future price movements using sentiment analysis on Twitter. The Tie describe their data as, “a quantified representation of investor intentions expressed on Twitter void of noise generated by spam accounts.”
According to The Tie, sentiment analysis has proven to be a powerful leading indicator for future cryptocurrency prices, and it has proven to be more effective…