How To Build A Document Annotation Project?


Tag and mark entities in PDF, DOC, DOCX or text documents.

A quick overview of the full process, uploading data, doing tagging and downloading results.

Uploading data

You can upload dataset to be tagged while creating the project and also add any additional data items to the project anytime you like.

There are two options of uploading data:

(1) You can upload a single text file (PDF, DOC, Text etc) where entire document becomes a data item to be tagged.

(2) You can upload a zip file of all the documents. The zip would be automatically extracted to get all the documents from it. Each document becomes an independent data item to be tagged.

Downloading results

The results are downloaded as a text file where each line is a JSON containing the text of the document and defination of each entity labeled in the document.

Example annotation:

"points" are each of the entities tagged in the document. "start" and "end" are the character-offset of the phrase starting from the begining of the document.

Privacy

For paid plans (even during trials) all data is completely private. Dataturks claims no ownership of any of your data.