Challenge background
Japan, like other developed countries, has a large population of elderly people living on their own. They do not always have relatives nearby or access to community care. Because of mobility problems and also events like heatwaves and COVID, they are reluctant to leave their homes, causing isolation and putting them at risk of health emergencies.
The problem
Elderly people who live on their own are at risk of suffering medical emergencies. Being on their own, their relatives or the health emergency services may not be alerted on time, increasing the risk. Inclement weather (e.g. heatwaves) increases the risk even further.
Using IoT (e.g. home energy consumption data) will help identify signs of inactivity, which, combined with other data (e.g type of housing) will allow designing an alert system for elderly people who may be in need of urgent attention.
Goal of the project
1. To find a reliable source of IoT streaming data (home electricity use, water use). 2. To determine a model that would work in this domain. 3. To build a model for individual health emergency risk that combines IoT with other data (e.g. type of housing, area indicators). 4. To build an interactive app that outputs alert levels for addresses with inactivity.
Project timeline
- 1
Week 1
– Domain Research
– IoT data and other data collection
– Tutorials
- 2
Week 2
– Exploratory Data Analysis (EDA)
– IoT data structuring
– Risk model building
- 3
Week 3
– Risk model building
– Tutorial (Streamlit)
– Dashboard building
- 4
Week 4
– Integrating work
– Deployment dashboard app
What you'll learn
Learn how to:
– Source and work with IoT data – Structure data and conduct exploratory data analysis – Build a scoring model for emergency health risk – Build interactive dashboards (Streamlit)