In this project, I applied text analysis on twitter to observe people’s reactions and emotions regarding to any news related to bitcoin (or other cryptocurrencies.) Since it is a less-than-one-month project and Twitter set the limit rate for retrieving data via API, the comparison only covered data of three weeks.
The generated wordclouds and word frequencies can be different because of different text cleaning methods. In this project, I tried two cleaning approaches. The below results were generated without moving hash tags(#) and @. As a result, more key words and user name remained while less emotional words were dispayed. To see other results by removing all hash tags, @ and bitcoin-related key words, please refer to Shiny-app version. In this version, more words other than market terms are presented. Some emotioal words can be found in those wordclouds.
The two peaks occurring at 4/10~4/11 and 4/22~4/24 are noteworthy. We can focus on these two dates in the below word frequency analysis and word clouds.
It is important to consider the whole market performance rather than only focus on bitcoin. Since people invest in other cryptocurrencies also have interest in bitcoin. All keywords are related.
##
no new words
From 4/24 to 4/27, neither many new words nor siginificant words appeared, which imply that nothing new happend during these days. And it may also correlated with the low volume of bitcoin trade at the same time.
no new or significant words
Even though all the tweets were retrieved using keywords “Bitcoin”, the texts were actaully not only dicussed about bitcoin but the whole cryptocurrencies market. For example, in 4/21, there were some positive words such as increase and earn. But the price and voulme of boitcoin actually didn’t have significant changes, while the total market cap excluding bitcoin increased on that day. And other cryptocurrencies such as altcoins, bitcoin cash, Ethereum and EOS were mentioned a lot during this time period on tweeter. As a result, we may infer that investors of bitcoin also concern and have interest in other cryptocurrencies. In addition, I personally think people rarely use many emotional words when discussing about cryptocurrencies on Twitter. Most words were neutral. And lots of market terms were extracted from the texts. Nevertheless, from all siginificant words of each day, we can depict a timeline marking all big news in this time period and see how people reacted to these news in terms of investment in cryptocurrencies. And if the timeline expand, we might get more insightful results, considering people’s different reaction times.