Usage statistics

Usage statistics can only be displayed by system administrators.

Usage statistics (for training and translation) are visible to system administrators, and accessible through Administration > Usage statistics.

The displayed character counts are:

  • The source + target characters in all corpora in an engine (excluding stock corpora) for training

  • The source characters in translatable segments for translation

Only successfully finished training/translation jobs generate statistical entries. The month into which an entry gets logged to depends on the finishing date and time of the job, in UTC.

Aggregated usage statistics

Statistics by month

The statistics dashboard shows all training and translation statistics up to now, with the current month highlighted.

By clicking on a month, you can access a detailed list of engines that were involved in either training or translation in the particular month.

Statistics by engine

If you are interested in which engines are used most, or in the usage history of a particular engine, go to the Engines tab. Here you will find the sum of characters trained and translated by all engines. By clicking on an engine, you can see the “activity” history of the particular engine over past months.

Detailed usage statistics

Detailed training statistics

Detailed training statistics can be displayed on the Training tab. You can

  • filter the displayed rows by engine name, languages and dates;

  • export training statistics into a CSV file with the current filters applied.

Detailed translation statistics

Detailed translation statistics can be displayed on the Translation tab. You can

  • filter the displayed rows by engine name, languages and dates;

  • export translation statistics into a CSV file with the current filters applied.