Skip to content
This repository has been archived by the owner on Apr 29, 2021. It is now read-only.

DVU Overview

Jump to bottom
Joshua Selsky edited this page Sep 16, 2013 · 2 revisions

Data Visualization Units take an Open mHealth data stream (identified by a schema ID and version) and display it in a meaningful way to an end user (a patient, clinician, caregiver).

To date, Open mHealth has built out dashboards using a kind of meta-DVU concept. Each dashboard pulls in multiple data streams to give an overall view of a patient's well-being with regards to a specific clinical domain.

Here is an example of a browser-based PTSD dashboard that is meant to be used during a clinician-patient visit.

PTSD View

The selections at the left control which data streams get displayed in the main graphical view on the right. Underneath the graph, aggregate details about a specific self-report item (in this case, sleep) are displayed for the selected range within the scrollable date slider.

Here is an example of a smartphone-based diabetes dashboard called My Comparisons that is meant to be used by a patient as they collect data.

Diabetes View

This dashboard pulls together data from the following sources: Entra glucometer, self-report on food intake using ohmage, RunKeeper self-report on exercise, and self-report on mood using PAM (Photographic Affect Meter). The idea is to compare nuances of different data streams to each other to answer questions related to well-being. This image shows how both high and normal blood glucose readings relate to food intake; how negative mood relates to high blood glucose readings; and how positive emotions relate to physical activity. The idea is that it is a living diary in a person's pocket to use for self-reflection. On the explore tab, the patient can add different comparisons to the main dashboard view. For more information, check out the My Comparisons wiki page.

Both of these examples pull together many mHealth data streams. One of the goals of the Open mHealth software architecture is to foster reusability of software components. This is where the idea of a DVU comes into play. We'd like to foster an open repository of software that could apply to many clinical domains. For example, the PTSD Dashboard could be used in other psychiatric domains and the diabetes dashboard could be used for pre-diabetes or generalized for basic activity tracking.

We are standardizing on HTML5 and Javascript as a baseline for DVUs. Other programming languages are certainly relevant here as well (Processing, ActionScript, R, Java (Android), Objective-C (iOS)).

The core requirements for DVUs is that they must document what schema ID-Version pairs they visualize and they must be open source. The goal is to add DVUs to http://registry.openmhealth.org as they are built.

If you have questions, please feel free to open an issue in this repository or join our Google Group.

Clone this wiki locally