Company Event Popularity for Financial Markets using Twitter and Sentiment Analysis
- Mariana Daniel, Rui Ferreira Neves, Nuno Horta
- Expert Systems with Applications, 2017
- A version of this paper can be found here
- Want to read our summaries of academic finance papers? Check out our Academic Research Insight category.
What are the research questions?
- Is it possible to develop a system for the detection and discovery of the popularity of special events on Twitter that may influence the financial markets?
What are the Academic Insights?
The system devised by the authors is composed of five steps:
- Identification of the financial community by identifying relevant users ( those with more followers than users followed) based on 10 influencers
- Filtering of the tweets based on Investopedia financial dictionary
- Definition of sentiment by using four text analysis tools ( MySentiment Api, TextBlob, SentiStrenght and Affin)
- Normalization by volume
- Event detection
The authors study over 12 million tweets from 9,000 users over 710 days ( from September 2013 and September 2015). Of these tweets, only 192,935 (those with content relevant to the financial markets) are analyzed.
The three case studies presented in the paper (Apple, Microsoft and Walmart) are examples of the efficiency of the model in dealing with massive amounts of data and very high levels of noise typical of social networks.
Why does it matter?
While this is not the typical paper that researches relations between data and market performance, it shows the details of how to approach sentiment analysis on Twitter starting from the definition of a relevant community of users down to the definition of a popular company event.
The Most Important Chart from the Paper:
The growing number of Twitter users makes it a valuable source of information to study what is happening right now. Users often use Twitter to report real-life events. Here we are only interested in following the financial community. This paper focuses on detecting events popularity through sentiment analysis of tweets published by the financial community on the Twitter universe. The detection of events popularity on Twitter makes this a non-trivial task due to noisy content that often are the tweets. This work aims to filter out all the noisy tweets in order to analyze only the tweets that influence the financial market, more specifically the thirty companies that compose the Dow Jones Average. To perform these tasks, in this paper it is proposed a methodology that starts from the financial community of Twitter and then filters the collected tweets, makes the sentiment analysis of the tweets and finally detects the important events in the life of companies.
- The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
- Join thousands of other readers and subscribe to our blog.
- This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.