Worldwide, there are more than 300 million depressive people according to the World Health Organization. Depression not only accounts for the majority of chronic sick leaves. It's one of the most frequently mentioned causes when it comes to the 800k suicides per year. MoodAmigo's mission is to support those who are currently undergoing treatment and have a hard time getting back to normal life during their rehabilitation.
What it does
MoodAmigo acts as a bridge between people who suffer from depression and their circle of friends. As soon as MoodAmigo detects that the patient should be cheered up by human interaction, it informs the specified acquaintances.
MoodAmigo consists of a bracelet that collects the following data and a companion web-based dashboard that displays its analysis.
- voice in order to detect the carrier's sentiment
- decline in physical activity
- prolonged insufficient exposure to light
The percentage number in the doughnut charts of the dashboard expresses how likely the detected respective mood is real. As feelings may overlap, the percentages don't add up to 1.
How we built it
Regarding the hardware, the mentioned features rely on a microphone (for recording the voice), a movement sensor (for recording physical activity), and an ambient light sensor (for recording insufficient exposure). All of them are included in the Simplelink CC1350 SensorTag by Texas Instruments, which is the core part of the bracelet. It is able to connect via Bluetooth and sub-GHz. Because the battery lifetime for operation on sub-GHz is longer, it is our preferred option.
The data is submitted to IBM Bluemix and acts as input in the Node-RED application. MoodAmigo analyses the data with several Watson services, such as the personality insights, the tone analyzer, and the sentiment analyzer. Dashboard information is continually updated and in case of a negative trend MoodAmigo messages predefined acquaintances.
To avoid sending out false alarms, the sensor data is aggregated over a period of time. For this purpose, a sliding window protocol is used.
Challenges we ran into
On the second hackathon day we discovered that there are no sound drivers yet implemented in the SensorTag. Consequently, MoodAmigo is temporarily relying on a smartphone application for voice input.
Accomplishments that we're proud of
As the developers of MoodAmigo we're proud of having achieved a step towards helping depressive people improve their life by using our service. MoodAmigo's unique benefit lies in acting as a gateway to successful depression treatment by involving family & friends.
What we learned
We learned how to build a working prototype with Bluemix, Node-RED, and the Texas Instruments SensorTag from scratch.
What's next for MoodAmigo
MoodAmigo's concept is easily scalable to support patients with other mental disorders, with physical rehabilitation, or simply elderly people who are living alone and need assistance.
Besides, there are many technical improvements required, which were out of scope of this hackathon, such as a database to store the individual records.