Increase alignment of i2b2, MCW and Mayo PHI datasets

[tschaffter]

We now have the Mayo Clinic dataset mapped to the NLP Sandbox schemas. We are currently working on connecting this dataset to the benchmarking infra of NLPSandbox.io.

Example of information to compare:

• Number of patients in the dataset
• Number of clinical notes
• Distribution of notes per patients
• Number of annotations
  • Number of annotations grouped by annotation type (date, person name, etc.)
• etc.

Tasks:

• Develop a notebook to compare the content of multiple datasets #22
• Document the difference between the i2b2, MCW, Mayo PHI datasets #24
• Update MCW dataset to improve the standardization of PHI datasets #25
• Update Mayo dataset to improve the standardization of PHI datasets #26