Automatic Brand Analyzer for Social Networks and Other Internet Resources

Project summary: the company has developed a search engine and a system for semantic analysis of hotel brands references. The system was integrated into a SaaS solution that enables hotel owners to manage their brand presence in Social Media. It searches for popular sites with customer reviews of hotels, such as booking.com, expedia, hotels.com, tripadvisor, zoover.de, hrs.com and others, as well as for comments in the social networks Facebook and Twitter. The system returns to the sites for updated information on a regular basis. For this purpose we implemented search robots (crawlers). Reviews' texts are analyzed with the help of a neural network (we applied Apache OpenNLP project) from different perspectives: negative or positive, comments about Wi-Fi, garage or room service and so on. The system composes a hotels' rating on the basis of the obtained information and allows the owners to review feedbacks collected from various sources and compare the ratings. All these helps them to come up with new ideas for service quality improvements.

Technologies: Java, IceFaces, Vaadin, MongoDB, Twitter/Facebook API, Spring (IoC, AoP, Batch), JUnit, Mockito, Apache Tomcat, Eclipse, Git, Maven, Ant, Gerrit (system for code review), Jenkins, Nexus, Apache OpenNLP, shell scripts, Web Driver (UI-tests), Quarz, RegExp.

Project features:

  • Full compliance with the Scrum processes as one of the customer's requirements.
  • Use of a Git/Gerrit combination.
  • One of the teams worked on the side of our customer.

Project results:

  • All iterations with the participation of our team were acknowledged successful
  • Information gathering was implemented for more than 25 sites with customer reviews of hotels.
  • The team made a business trip to the customer's site in Germany, experience exchange with the colleagues and delicious German beer rendered the trip successful 🙂

Company’s achievements during the project:

  • The project team developed a number of architectural improvements for the crawlers implementation, such as data gathering on different language versions of a site.
  • There was implemented a system for regressive crawlers testing, which uses the data collected by the crawlers.
  • We implemented an automated UI testing using Web Driver, and created an architectural framework for auto tests using the Page Object pattern.

Testimonial:

Friedrich Boeckh
Friedrich BoeckhManaging Director, Infopulse GmbH, Information Technology and Services, Germany

JazzTeam participated in innovative project I was supervising.
A team shown high professional level in a matter of architectural design, Agile processes, unit testing and involvment at the project.
You could trust JazzTeam engineers, rely on them and could expect high quality provided by Java experts.