Human Experience Management System is a knowledge management application which combines traditional software development and data-mining techniques. As a result an application was made which uses company’s internal data from project management systems and external data from the internet (LinkedIn). Project data show us the professional track, the experiences of an employee which can be used to determine the expertise of the specific employee. This in particularly applies to LinkedIn profiles where people create their online CVs to show their expertise and this gives a good quality external source for staffing and recruitment.
The application aims to facilitate the daily work of project managers and HR experts via the whole lifecycle of the project from staffing, via load balancing of the employees, to end of the project evaluation. It holds five main functional areas:
1. Browsing of database’s data in a tabular format
2. Visualized work-load distribution of employees regarding both running and closed projects
3. Text-based performance evaluation possibility for managers after ending of the project
4. Keyword based expert finder for helping the staffing of a project with the usage of text-mining technologies
5. LinkedIn integration:
a. Seniority comparator based on the spent years in work
b. LinkedIn profile importer tool
Technical terms speaking the project implements a JavaFX based UI (non-web, but Swing-like) which reads and writes data into the MySQL database via Java Persistence API. The data from the database (which can contain both internal employee’s data or data crawled from LinkedIn) gives an input for the data-mining process which was modelled in RapidMiner. The RapidMiner process was created through RapidMiner’s GUI and manipulated programmatically with the input and parameters from the JavaFX based application. The results of RapidMiner are driven back to the JavaFX UI. Both the decision support charts and tabular displays are programmed with JavaFX and FXML technologies.
In the light of aforementioned functionalities the whole task has been completely implemented and tested. The testing was done both as an automated process (JUnit) and as a manual system test. The demands articulated in the thesis assignment were fully accomplished and on some points even overperformed.