LaVA corpus is published in Bonito *.vert file format with error and morphological annotations. The configuration file for SketchEngine is also available. These files are already uploaded to a freely available instance of noSketchEngine.
If you want to take advantage of the additional error analysis capabilities provided only by the paid SketchEngine version, you can upload the vert file with the appropriate configuration file in your SketchEngine profile.
All corpus data is available for download and further processing in two csv files.
File essays.csv contains meta data, original and corrected data.
File annotations.csv contains token level alignments between original and corrected text with all annotations layers - manual lemmatization, manual morphological annotations and automatical error type classification.