If the dataset meta data is also stored, have some stats present about how far the annotation is done - What percentage is unlabeled/labeled - Pie chart, how many (of total/ of labeled) data entries are labeled.