Sensor data

Starting point

• Automotive manufacturer uses crowd sourcing to record sensor data from test drives
• Many hundreds of millions of data sets are available
• Data sets are not directly interpreted by the engineers

Objectives

• Interactive data exploration
• Temporal, spatial and content-related selection for drill-down
• Data should be used to understand the influence of various factors on the behavior of driver and vehicle
• Data exploration should form the basis for further development of services like assistance systems

Course of action

• Requirements analysis together with specialist departments
• Recognition that available tools cannot fulfill the requirements leads to native development
• Apache Solr as back-end component allows very fast access to geo-regions, times and sensor data
• Frontend components: Maps with Leaflet (OSM), visualizations and statistics with D3.js, Angular JS as application framework

Results and benefits

• Intuitively usable interactive application for data exploration
• The use of scalable open source technologies ensures real time capability – also in connection with dynamically growing data pools
• Researchers are able to recognize patterns visually as well as to investigate correlations and patterns conveniently