Explore your data

Starting with a blank notebook, this page walks you through connecting to the JupyterHealth Exchange and viewing the data you can access. It assumes you have logged in and launched a session.

Connect to the Exchange¶

Your Hub session provides two environment variables so the client can reach the Exchange using your account, without asking you to log in again:

The client library reads these automatically.

Open a new notebook with File › New › Notebook and run the following code blocks in order:

from jupyterhealth_client import JupyterHealthClient

client = JupyterHealthClient()

Tokens expire after a day. If requests start failing, log out, log back in, and restart your server.

To use the client outside the Hub, get your own token.

Check your connection¶

Run this to see your user information:

client.get_user()

List studies and patients¶

Data is organized by organization, study, and patient. List the studies you can access. You’ll use one of these IDs in the next step:

print("All my studies:")
for study in client.list_studies():
    print(f"  - [Study ID: {study['id']}] {study['name']} (org: {study['organization']['name']})")

Choose an ID from the list above and replace <STUDY_ID> before running this cell:

study_id = <STUDY_ID>

client.get_study(study_id)

List the patients in that study:

print("Patients in this study:")
for patient in client.list_patients(study_id=study_id):
    print(f"  - [Patient ID: {patient['id']}] {patient['nameGiven']} {patient['nameFamily']}")

View patient data¶

Choose a patient ID from the list above and replace <PATIENT_ID> before running this cell. This example retrieves blood glucose observations:

from jupyterhealth_client import Code

patient_id = <PATIENT_ID>

df = client.list_observations_df(
    patient_id=patient_id,
    study_id=study_id,
    code=Code.BLOOD_GLUCOSE
)
df.head()

If the DataFrame is empty, this patient may not have blood glucose data available for this study. Check their consent with client.get_patient_consents(patient_id=patient_id), or try another patient or data code.

For an example of glucose metrics and plots, see the CGM tutorial. The study and patient IDs in that tutorial are illustrative, so its code won’t run as written. Replace them with IDs from your own study. The client API reference lists the other data codes and methods.