That helps Aditya. Thanks!
One suggestion, while waiting to load data, it may be nice to show loading please wait status instead of showing blank page until data gets loaded.
1 Like
Hi Aditya,
Even after we increased the RAM size in our CentOS server, we are noticing mindLAMP dashboard application is slow in response to view the data.
I am accessing the mindLAMP app from within corporate network using the actual server name as the domain name in the mindLAMP dashboard using Chrome Browser. I have high performance computer with good network speed.
Some of the requests takes about 12 - 16 seconds to load data and view the page content, while looking at the browser network log most of the time they are waiting for attachment API requests. I have attached screenshot from Chrome browser.
Please let us know if there is a way to reduce the page load time. Thanks!
@jeyraman Could you take a look at whether CouchDB has the appropriate design documents created for the tag database? This may be related to your previous issue as the database may not have fully initialized for some reason.
Thanks Aditya for your response. I see one design document for tag database attached the screenshot here.
I do not see any design document for sensor_spec and activity_spec.
@jeyraman Could you make a GitHub issue so we can follow up there? Please be sure to include this information, the screenshot, as well as console logs from the API Server as well. Thank you!
Sure, I will create GitHub issue with screenshots and logs from browser and API server, Thanks a lot!
1 Like
Hi Aditya, We like to see data from Smart watch such as step count/heart rate etc.
One of the project team member has installed the mindLAMP app in Smartwatch and logged into the Smart watch mindLAMP app, but we are not seeing sensor data in dashboard. Where do I see them? Do we need to enable any thing to get sensor data from Smart Watch? No data in Cortex either. How to get Cortex data.
In the CouchDB database, I do not see any other sensor data except 3 design documents. But we have lot of sensor event data.
Here is the screenshot of the activities.
Please let us know if any configuration missing for collecting smart watch sensor data. Thanks!
@jeyraman Could you clarify which database in CouchDB you were checking (sensor_spec, sensor, or sensor_event)? Furthermore, please check incoming data using the API:
curl \
https://<your_server_address_here>/participant/<your_participant_id_here>/sensor_event \
-H 'Authorization: Basic admin:<your_admin_password_here>'
If you’re using the Watch apps, you may see lamp.watch.accelerometer instead of lamp.accelerometer. This is a bug that will be rectified in the upcoming release (later this month).
Hope this helps.
Thanks for your response Aditya.
When I execute above curl command I see lots of data from multiple sensors (lamp.guyroscope, lamp.magnetometer,lamp.accelerometer.motion etc ) from sensor_event database.
Here is some:
{
“timestamp”: 1612884145932,
“sensor”: “lamp.magnetometer”,
“data”: {
“x”: 6.407135009765625,
“y”: -18.130661010742188,
“z”: -80.86616516113281
}
},
{
“timestamp”: 1612884145932,
“sensor”: “lamp.gyroscope”,
“data”: {
“x”: -0.0032740901224315166,
“y”: 0.008932508528232574,
“z”: -0.00917299184948206
}
}
No data with sensor name lamp.watch.xxxxxx.
How to see those sensor data in the dashboard app?
Hi @jeyraman,
The sensor_event database will accumulate a vast amount of data over time, even for a single participant and so it is not feasible to stream tens or hundreds of gigabytes of data to the web dashboard or mobile app. For this reason, it is not possible to view raw sensor data (such as lamp.accelerometer, at over 5 Hz frequency) in the UI, and you will need to use the Cortex data analysis/processing toolkit to produce summary metrics and secondary features that are then easily visualized in the UI. This is something that can be done by the researcher or data scientist on your team as Cortex is provided as a publicly-available Python package. For more information on Cortex, please consult our documentation.
Hope this helps!
Thanks Aditya for your response. I will look into Cortex documentation.
I see following sensors data is getting collected - lamp.accelerometer.motion, lamp.accelerometer, lamp.analytics, lamp.gyroscope, lamp.gps, lamp.magnetometer
Do you collect heart rate and step count etc from the smart watches?
Or we need to use accelerometer, motion data to find out step count.
Thanks!
We collect both raw sensor data (accelerometer, etc) that can be used to compute step count manually as well as HealthKit/GoogleFit-provided sensor data (aggregated pedometer-based step count) that provides this data already out of the box.
1 Like
Thanks for the information and immediate response!
1 Like
If the person has a wearable that syncs to Apple Health or Google Fit, then that wearable data should be accesible too via this mechanism.
1 Like
We like to see these sensor summary data in our dashboard. Is there anything we need to do to enable this?
Can you let us know how frequently the sensor data gets synched from Apple Health/Google Fit to mindLAMP database?
@jeyraman You unfortunately will not be able to see raw sensor data in the dashboard due to the large volume of data that will need to be streamed to the browser (and may crash the app/browser). You will need to run the Cortex pipeline and generate graphs that are then saved and displayed here. Hope this helps.