Visual Analysis of Financial Reports

Type: 
DA
Persons: 
1

Description

In the financial domain, trend analysts have to read many finance reports to be able to foresee financial developments. However, to be able to make well-informed decisions, analysts have to consume a lot of textual information, which is time-consuming and error-prone.

Tasks

To ease this process, this project should provide a compact visual summary of the temporal development of these reports with respect to automatically extracted, as well as manually defined topics.  In addition, trend analysts should have the possibility to train the system whether the information contained in the reports is considered to be positive or negative. In the end, the system should visually summarize the latest subscribed RSS feeds for the trend analyst, so that she can efficiently decide which kind of information is contained, whether this information is positive or negative for the market, and how this topics and their implications for the market developed over time.

Requirements

  • Strong interest in information visualization, machine learning, and natural language processing.
  • Experience in web programming, in particular JavaScript.
  • Experience with web technologies, like Node.js, Angular.js, or d3.js, advantageous.
  • Commitment to a tight collaboration with domain experts from the financial domain.
  • Knowledge in the financial domain is not required, but the student should be interested in the topic.

Environment

The project will be built upon an existing online financial dashboard system, using Node.js, Angular.js, d3.js, and Reactive.js. Within this system, there is already an existing report analyzer that incorporates common natural language processing techniques on RSS feeds, like part-of-speech tagging, topic modeling, or basic sentiment analysis (using JavaScript libraries, like compendium.js).

The work will be done in tight collaboration with an IT-company focusing on financial trend analysis (PS Quant: http://www.ps-quant.com/cms/), as well as scientific collaborator in the area of machine learning from the FH St.Pölten. In case of a successful completion of the project, there will be a monetary reward.

Contact

For more information please contact Michael Pühringer or Manuela Waldner (waldner@cg.tuwien.ac.at).