The aim of the project is to visualize typical established patterns of behavior by users of a software system used by a customer. To achieve this, analytics software is developed to graphically format the clickstream data and form the basis for discussing future optimizations.
Supplement
To perform the clickstream analysis, a proof-of-concept prototype is first created and various database systems (InfluxDB, MySQL) are evaluated. After this, the web interface can be designed on the basis of the customer requirements using Adobe Fireworks. In order to utilize the benefits of single-page applications (SPA) when developing the web front end, the widely used SPA framework AngularJS is to be used. To present the data analytically, dynamic and interactive graphs are generated using the JavaScript library D3.js. The actual evaluation of the clickstream raw data is implemented using a performance-optimized server application written in C++. Thanks to efficient algorithms, the resultant in-memory solution can process several million data records promptly. To keep ongoing operating costs as low as possible, all the components are delivered as a standalone solution in a software bundle.
Subject description
Definition of evaluation scenarios and interfaces for answering the important behavior questions: What are the top 5 patterns of behavior when using the application? What does the distribution of usage durations look like? What is the average time spent on the overall application and the individual components of the application?