The Sentinel project provides a dashboard presenting insights into the volume and quality of the conversation towards followed brand, products or services in a time-aware manner. The most important features include:
- Tracking keywords appearances - real-time monitoring of occurrences of selected keywords or their variations on the Internet. Keywords can include, but are not limited to: brand name, key products or events, most important staff members, company-owned domains. User can manage followed keywords using the website.
- Evaluating sentiment expressed in the text - numerical summary of emotions taking into account keywords and evaluated automatically.
- Listing of mentions - direct links and content of all of the mentions; to allow for quick response from the PR department.
Data streams consumed in this project come from multiple sources: Google News, Hacker News, Reddit and Twitter. Diagram presented below shows the general flow of the data throughout the system:
Multiple services, based on the AWS platform, were used to ensure high scalability of the system. Most of them are listed below:
Comprehensive description of the project (including architecture) can be found in report #1 and report #2.
The project was developed for simultaneously Big Data Analytics: Methods and Systems and Cloud Computing classes at the Warsaw University of Technology by Marcin Dardziński, Paweł Rzepiński, Ryszard Szymański and Sebastian Sowik.