Learn What a Behaviorgram Is and How It Works
We have a talented team of data scientists and researchers focused on providing individuals the opportunity to participate in cutting edge research. When our members participate in research, the activity data that they consent for use in the research helps our team study a number of health behaviors — everything from how to measure chronic pain to better understanding the individual migraine experience.
Analyzing and contextualizing the behaviors of thousands of participants with millions of data points is not an easy task. Over the years, this analysis lead to the creation of something that we have internally named “behaviorgram”. The behaviorgram of a person is a collection of their data streams over the course of some time. A behaviorgram can be quite simple and consist of only a few data streams from a single device (such as heart rate and steps from an activity tracker) or it can be quite large and include multiple data streams from multiple sources. A behaviorgram can be visualized much like an ECG, and below, we provide a glimpse as to what this visual may look like.
In our studies, we may start with activity data from a Fitbit or Apple Watch and then layer on additional data such as survey responses (e.g., mood or feelings during the day) and contextual information (e.g., weather, air quality, and the time of the day). Dense behaviorgrams give us richer pictures of someone’s behavior over time. As a reference, the behaviorgrams we created for the research on Alzheimer’s disease we did with Eli Lilly and Apple in August contained dozens of data streams and included steps counts, walking pace, sleep depth, workout sessions, and many others.
Behaviorgrams help us put data into context and extract additional insights from it. For example, we can more easily analyze how heart rate changes when people are in a workout session, or the effect that talking on the phone has on walking pace. It also allows us to derive new data channels from the existing ones. For instance, while MyFitnessPal already reports how many grams of protein and calories are eaten, we often find it useful to layer in a channel that represents the percentage of calories that came from protein. Similarly, we often add a new data stream that represents heart rate during periods of inactivity (i.e., no steps).
What would actually help you visualize a behaviorgram? A picture of course! We wanted to illustrate that there are small moments throughout your daily life that can have a significant impact on your health, and we were fortunate enough to get consent to share one of our colleague’s visual health journey via the various apps they have connected on Evidation. Looking at their data from the last 6 months, we tracked down a few key dates where we noticed specific behavior changes and asked them to add labels to the behaviorgram on what happened in their life on those days. Sure enough — in a week with less sleep, we uncovered they were abroad, working until late to match California time and then having troubles falling asleep due to jet lag and an uncooperative baby. You can see other activity trends like sightseeing with friends in Singapore and New York on high steps count days, frequent awakenings in sleep during a night spent on an international flight, and a spike in logged calories when they had leftover sweets for breakfast. You can also see that they were able to sleep longer when the baby was away with the grandmother. We got a good laugh out of the observation that the company offsite was a high fat/protein-heavy day!
Based on weather data, the highest precipitation day is low in activity! Though some of our team is based in sunny locations, it highlights that weather does impact our activity. Related, given the air quality impact that the Wildfires had in California this year, you can see that during the time of the fires 11/25 to 12/14, they did not spend much time outside and took some time off from exercising (with the exception of their time in New York!)
Keep an eye out for more references to behaviorgrams from us! Our hope is that they’ll begin to become a tool by which you contextualize different actions that you take. We also know that big life changes can impact your habits and health. That’s why we’re planning on sharing some personal stories from our team on the journey to being a new parent and its impact on health routine. We also want to capture events like life with a newborn to highlight trends that our specific team members are experiencing to show activity patterns during the first few months of these very significant milestones.