Financial analysis and forecasting is one of the spheres of human activity that has benefited a lot from the introduction of modern technologies. However, this process was not quite a smooth one as it required a number of important changes that transformed the field and made specialists adapt to the new conditions. First and foremost, there appeared a new source of financial information besides business publications, scholarly research, and financial reports of business organizations. Different stakeholders now have access to a number of valuable digital resources.
We will write a custom Essay on Financial Analysis and Forecasting Using Twitter specifically for you
301 certified writers online
First and foremost, the Internet and various new portals provide relevant financial reports, which is especially important for investors as they have to be aware of all the changes that happen in the market and can affect their practice. Furthermore, it would be fair to claim that such sources are more valuable than paper reports since they are updated regularly and contain data that is still up to date and can influence the decision-making process of various businesses.
There exist social media resources that allow analyzing the choices of consumers and market fluctuations in connection to social trends. This is especially helpful in financial forecasting as consumption choices are usually irrational and have to be studied in order to understand what ideas customers hold about different purchases, industries, brands, prices, and other things that make them opt for one product or service while neglecting the other.
We have explored how the financial data found on the famous social network Twitter can be used by a financial forecaster for creating the reliable and valuable basis for financial decision making and identified what particular benefits the network has. We discovered that density of messages makes postings useful and effective and they provide dynamic financial information free of redundancy.
The research was particularly focused on timing as the most essential aspect of financial operations. All financial players strive to complete their operations in the most accurate and beneficial time, which means that they need data as soon as it appears. When professionals make use of online resources, it is necessary to gather and process large quantities of incoming information in order to extract the most valuable facts.
Semantic frames that underlie textual data have to be identified to understand the major concepts. Social networks are particularly important in this aspect as they serve as massive sources of information and provide quick updates that are necessary for financial forecasting. In Twitter, a specialist can identify users whose opinions are the most accurate and reflect all the changes in the stock market. With the help of Latent Dirichlet Allocation (LDA), a researcher can process a series of determined topics and compare them to the sentiment in the shared Twitter content.
In this research, we collected data collected from Twitter and Google in order to prove that Twitter data has more predictive power about financial stock markets than blogs and financial news. We conducted three searches including
- top gainers
- top losers, and
- general search on stocks.
We collected the data of almost 10 000 tweets from January 1st, 2017. Twitter corpus was collected using Python program bypassing Twitter API limitations.
We found out that most of the tweets are irrelevant and contained unnecessary information, which meant that filtering was required. Using UNIX text tools, we removed URLs, retweets, usernames and stopwords and applied sentiment analysis. It was discovered that in comparison to Google, tweeters are less likely to tweet about losses than gains. There are not many negative words as well as words for highest losers’ words, almost nothing.
There are two sections that are words that appeared in tweets about ”Gainers” and “Losers” companies. In “Ganiers” graph has words with positive and neutral scores and words with only positive scores. Words with positive and neutral scores clearly shows that there are more positive words and the sentiment scores are much higher.
We also found out that companies that are highest gainers and losers in stock and with highest rising stocks are companies with technology, which conveys a societal trend at the moment.