In 2015, Robert Wood Johnson Foundation funded a pilot study to look at new ways of measuring family caregiving. Family caregiving is often overlooked—and rarely studied—in part because it’s difficult to measure. We set out to collect measurements, and make an atlas of caregiving, or at least a prototype. The key idea was to learn: How might IoT sensors augment traditional ethnography in learning about family caregiving?

We looked at 14 households, with 20 participants, including over 21 chronic illnesses. We used 12 sensors, measuring 16 factors, over an average of 24 hours, resulting in 5 GB of data (just from 1 sensor).

Processing the data required special knowledge and special tools.

Data Wrangling
Wrangling this amount of data was a challenge.
Spreadsheets for participants typically contained 21 million rows of data.
Simply finding tools that could process the data was a challenge.

Down Sampling
We collected more data than we could ever hope to show, so it had to be down sampled.
There’s a variety of methods for down sampling, each with their own pros and cons.
We contacted a math specialist to help us explore the options and understand the tradeoffs.

Filtering
In addition, some of the data required filtering.
There’s many ways to filter data.
Again, we contacted our math specialist to help us in this area.

Design projects increasingly involve teams. This project included help from fifteen people, and required a wide range of tools.

Healthcare—and the rest of our lives—are about to experience an explosion of sensor-generated data.

We see this project as a single of things to come. All designers will need to be well-versed in the wrangling, down sampling, and filtering of large data sets.

Many documents were created as part of this project:
Study Report
Data Visualizations
Diagram Development
Care Network Rationale

Learn more about the project by viewing our presentation or visit the Atlas of Caregiving website.