How is aggregating performed?
Once the usage data is harvested in OpenAIRE, it is cleaned and harmonized to procude more accurate statistics. Some of these transformation processes include:
- Removal of all malformed usage events .
- Unique publication identifier generation (currently by cross referencing DRIVER's identifiers).
- Filtering out of all robot initiated requests(COUNTER and custom black robot identification lists).
- Multiple simultaneous streams consolidation.
- Type of request (download full text or view metadata) deduction for uknown types of events.
- Double click filtering according to the COUNTER rules.
- Unique requester detection, which uses multiple input event fields to enhance precision.
- Grouping of multiple events of each requester into sessions (inactivity rule of 30 mins).
When sufficient usage data is available to OpenAIRE, additional types of analyses will be explored in search of usage impact metrics. Also, when combined with other types of data (e.g., author ids, citation, etc.) additional results may be deduced.