4.2.4 Aggregated data

Using relations between segments, reduced data (e.g., segment attributes) are typically aggregated into smaller, more usable tables, suitable for data analysis or data interpretation. For instance, driver characteristics can be grouped together with attributes from one type of situation, to evaluate the impact of drivers’ characteristics on their behaviour in that situation. Reducing the resolution or down-sampling the frequency of time series data can omit sensitive and confidential details about the system while still capturing key patterns and trends of the original data.

The data resulting from aggregating different kinds of reduced data together are called aggregated data. Although they are generally linked to a specific research question, the aggregated data may be re-used with different statistical methods, or re-aggregated with other data, to quickly answer new questions without the need to go back to harder-to-use, raw data. As they don’t contain instantaneous values, aggregated data don’t allow potentially problematic re-use, such as pinpointing illegal behaviour from one specific driver or benchmarking a driving assistance system without authorization from its supplier. As a result, aggregated data are generally easier to share than other categories of data.