Visualizing Health Data

Editor's Notes

• #4: Living healthier lives is no longer only a nice-to-have for us as a country or as individuals, but increasingly a must-have.
• #5: We have all heard about the rising healthcare costs in the US. And even though the increase appears to be slowing down, the US still spends about two times as much as France, a country known for having great health services. Why are the costs so high? There are many reasons, but they likely include the rise of chronic illness, and a healthcare system that is optimized for treatment of disease and injury.
• #6: The statistics speak for themselves. One third of US adults and children are overweight, and another third obese. Only 5% of adults get the minimum amount of daily physical activity. This is a central reason our healthcare costs are so high. So what can we do?
• #7: Changing our personal behaviors is the best chance we have for improving our health. It is generally accepted that prevention is the key for both better health, and to lower healthcare costs. Prevention is cheaper than treatment—both for the individual, and for society as a whole.
• #8: But as we all know, staying healthy is hard. Busy lifestyles and other factors can get in the way of exercise and a nutritious diet.
• #9: The key to achieving better health lies in two primary factors. Measurability is about being able to measure your actual behaviors. This eliminates the guesswork and allows you to see whether you are actually improving. People generally have only a vague sense of their health and whether they are doing better or worse. Once you know the facts, you can live by them. Accountability is about being held responsible for achieving better health. Some manage to hold themselves accountable. But for those of us who need more support, people can be great motivators, for example your friends, or a personal trainer.
• #10: Recently I met Matt. Matt works in computers, and, like most of us, he is a busy individual. Still, he manages to find time to go running with his friends. However, because his friends all live in other cities, they compete with each other by logging their total miles and calories burned on a shared spreadsheet. And as a bonus, they track their miles against distances between cities, for example Seattle to Vancouver, or Seattle to Los Angeles, or the Earth’s Circumference. This is an example of measurability and accountability in action, helping Matt and his friends live a healthier lifestyle.
• #11: Matt’s story is also emblematic of a recent movement called the Quantified Self, which brings together people who quantify and track their own behaviors: Self-trackers. The Quantified Self movement began with a website created by Gary Wolf and Kevin Kelly, both editors at Wired Magazine, in 2006. They began holding regular meetings with people who were running interesting personal data projects.
• #12: Since then, the Quantified Self community has grown to span 162 Meetup groups world-wide and almost 30,000 members. The size of the community is proof that there is mainstream appeal to the notion of self-tracking.
• #13: At its core, self-tracking is all about behavior change. The cycle of behavior change can be broken down into four phases: 1. Measure is about data collection. 2. Visualize is about giving form to data that are inherently abstract. 3. Interpret is about drawing actionable conclusions from the data. 4. Act is about acting in accordance to one’s conclusions. This cycle is repeated continuously to measure and evaluate the cause and effect of one’s actions.
• #14: Data collection is a key part of the process. Until recently, it was difficult to track ourselves. Our only method would have been to write down our behaviors manually, making analysis and drawing conclusions difficult.
• #15: Three factors have made data collection more attractive for mainstream audiences. The first is automation. Today, most of us carry smartphones, and wearable devices are starting to find adoption. While athletes and others have used pedometers, heart rate monitors and other specialized tracking devices for years, these devices are making it possible for a much larger audience to collect data about themselves automatically. The second is portability. Once you have your data, you want to do something with it. Part of that includes porting it to other applications that can help you make sense of it, or combine it with other datasets. This increases the possibilities of what you can do with your data. The third is sharing. Social networks like Facebook and Twitter have made it acceptable to share the most mundane details of our lives with our friends and colleagues. And this has also increased the appetite for sharing—and collecting—data.
• #16: Moves is a smartphone app that tracks your activities. Leveraging the GPS and accelerometer in your smartphone, it quietly runs in the background, tracking your physical activity throughout the day. From the accelerometer readings it knows to distinguish different types of activities, such as running, walking and cycling.
• #17: Another recent app for data collection is Reporter, created by designer Nicholas Felton. While Moves is focused on physical activity, Reporter records other forms of behavior that would normally be more difficult to capture. For instance, it tracks the people you are with, how well you slept, whether you are working, and how many coffees you have had, among other polls. You can also create your own. The app alerts you at fixed intervals throughout the day that it is time to report, interrupting whatever you are doing to enter a few quick data points. These can then be analyzed later to draw conclusions about your behaviors.
• #18: In addition to apps, wearable devices have started finding adoption with consumers. Fitbit offers a range devices you can wear on your wrist that track your steps, count calories, and measure sleep based on how much you toss and turn at night. Like many other devices in this category, Fitbit offers a lightweight, unobtrusive method for automatically collecting personal data.
• #19: Other emerging technologies like Google Glass aim to make augmented reality mainstream. While there are many potential applications for Glass, it is easy to see how Glass could also make data collection even more effortless. Information projected onto your field of vision could also give you real-time feedback and allow you to see the cause and effect of your actions in real-time.
• #20: Data collection is only the beginning. The promise of Quantified Self is that you can interpret your data and draw meaningful conclusions, that can allow your to change your behavior in the long run. That is where data visualization comes in.
• #22: This is a well-known definition of data visualization from the book Readings in Information Visualization by Card, Mackinlay and Shneiderman.
• #23: Data visualization is a science and a technology, and art. It is by definition multidisciplinary, drawing from the disciplines of Computer Science, Human-Computer Interaction, and Graphic Design.
• #24: Jarke van Wijk offers a framework for how these three areas are related. The data that inform a visualization are collected, parsed, and filtered via scientific methods. Graphic design is used for representing the data. And through Human-Computer Interaction, people can interact with the visualization. As the diagram shows, visualization both interprets reality, but can also have an effect on it.
• #25: It is also useful to point out the differences between data visualization and infographics, which have become popular as a means for telling a story through diagrams and images.
• #26: This diagram can also help contextualize the role of data visualization. It shows a hierarchy of visual understanding. Data visualization lies at the threshold of data and information, while information design lies at the threshold of information and knowledge.
• #27: The benefits of data visualization can be thought of as three C’s: Context, Comparison and Control.
• #28: Context is about stepping back and seeing the big picture. It is about understanding the boundaries of a dataset, as well as the gaps and outliers.
• #29: The Opte Project was a visualization of the entire Internet in 2003. Although it shows millions of connected devices and servers, it is remarkable how small and tangible the Internet appears. Zooming out to see the boundaries of a dataset can make information feel more manageable.
• #30: This visualization shows 500 million connected Facebook users world-wide. It shows the areas of high and low density in countries across the world. Apart from areas of high concentration of Facebook users, the areas that are underrepresented or not represented at all are equally interesting. These are the gaps, and outliers are data points that are separated from the main system.
• #31: Comparison is about seeing patterns in the data by comparing individual data points with one another.
• #32: In order to compare information, it has to be organized. LATCH was coined by Richard Saul Wurman, the founder of the TED conference, who realized there were only five ways of organizing information.
• #34: Network might be a possible 6th method of organizing information through connections to other information. LATCH is an important framework when it comes to creating comparisons.
• #35: Control is about the ability to manipulate data in order to find stories or insights. Control comes from the ability to filter, sort or group data, as well as selecting individual data points.
• #36: SandDance is a project by Steven Drucker, Principal Researcher at Microsoft. It visualizes hundreds of thousands of data points and supports natural user interactions for sorting, filtering and grouping content. This example shows 50,000 census tracts across the country. They can be visualized by region, or geographically. You can also apply color, for example to show per capita income, and different ranges can be specified to filter the data. Sanddance shows how a visualization can be used to tell many stories from a single set of data, by giving people control over the data.
• #37: Data on their own are abstract—they have no form. In order to visualize data, we need a visual metaphor. A common metaphor we are all familiar with is the desktop metaphor we use to interact with our computers.
• #38: Metaphors can provide a framework for representing and understanding personal data.
• #39: Dashboards are a common metaphor for personal data visualizations. The purpose of a dashboard is to manage your data.
• #40: The Fitbit dashboard is a model used by many other services as well. Data categories are shown as tiles that can be organized in any order. They give you an overview of how you are doing.
• #41: Each tile leads to a more in-depth report, showing either activity, food consumption or sleep. Benchmarks help people understand how they rank relative to the larger population.
• #42: Fitbit also allows for setting goals that people can measure themselves against.
• #43: Dashboards can also be expressive. Mem:o is a data visualization app for the iPad. It lets you enter simple notes with quantitative values attached, for example how many calories you ate for breakfast. These notes are visualized as circles of varying colors and sizes. The visuals are optimized to allow people to pick up on habits and trends, and for that the app has a calendar view that allows viewing data over time. The app’s look was inspired by the graphic design style of the Werkplaats Typografie, a design program in the Netherlands.
• #44: Moodstats is an application that allows you to quickly rate how your day was, in six different categories. Using color-coded graphs and statistics you can see exactly how your moods have changed over the last week, month, or year.
• #45: Another common metaphor are traces. The purpose of using traces as a metaphor is to represent a record of an activity.
• #46: Aaron Parecki tracked his location using his smartphone over three years. This is a map of Portland, shown only by the routes that Aaron traveled. The colors represent each year.
• #47: In a project called Biomapping, the artist Christian Nold measured location and emotional arousal using a device with a built-in GPS that records the wearer’s Galvanic Skin Response.
• #48: The artist worked with groups of volunteers to cover cities such as San Francisco and Paris. The result were printed “emotion maps”, showing the data traces of the participants, along with their annotations.
• #49: Nathan Yau’s Where People Run shows the potential for aggregating personal data from many individuals into a single visualization. Populated with public data from RunKeeper, the visualization shows the areas in cities where people choose to run, along with the density of how popular an area is.
• #50: Stories are another common metaphor for visualizing personal data.
• #51: In Quantified Breakup, a journalist documents the weeks and months after her divorce in charts and graphs. These display her irregular sleep patterns, texting habits, and public displays of emotion. Here, data visualization is used for personal expression and as a form of emotional self-help.
• #52: Jonathan Harris’ The Whale Hunt is a photo-based visualization documenting an Eskimo whale hunt, based on images he took over the course of nine days. Images were taken at regular but changing intervals. Harris correlated the frequency of the interval to the intensity of the moment: the more intense, the higher the number of images taken.
• #53: The moments of high and low intensity are especially visible in a timeline view, where each column represents a 30 minute period.
• #54: An automatic slideshow view is perhaps the most effective way to experience the piece. The speed of the slideshow is correlated to the intensity of the moment, allowing you to re-experience the event.
• #55: Autographer is a device that allows telling your own quantified stories through imagery. A camera you can wear, it records images automatically at a specified intervals. You can later find the perfect photo from a large collection of images, or playback the entire stream to replay an event or an entire day.
• #56: Portraits are a metaphor used to convey individual identity through data.
• #57: Since 2005, designer Nick Felton has created the Feltron Report, a year-end summary of his life, showing everything he did throughout the year through data. Felton generates many of the graphs in his reports from data logged on mobile devices and visualized using Processing.
• #59: Stephen Wolfram, the founder of Mathematica, has one of the world’s largest collections of personal data spanning over 20 years. Wolfram uses automated programs to record his email, calendar appointments, keystrokes, phone calls, and physical activity.
• #60: Nike’s Year in NikeFuel allows FuelBand users to display an entire year’s worth of activity data on a single poster. Graphs are layered on top of each other, each one representing a day, and the height of the graphs and the colors show the intensity of the workout.
• #61: Each visualization becomes a unique identifier for an individual and their data, allowing you can see how you did compared to others.
• #62: The Quantified Self ultimately gives people access to the kind of analysis that you could previously only get in a lab. These data can also be used to drive larger research efforts.
• #63: Mapping Health is a project Schema worked on with the Center for Public Health Nutrition and the Urban Form Lab at the University of Washington. The project visualizes data from the Seattle Obesity Study 2, led by Prof. Adam Drewnowski, PhD and funded by an NIH grant. The study tracked 500 participants using GPS sensors over the course of a week each. The project resulted in two videos.
• #64: The first video is called Slow/Fast Seattle. It shows GPS activity for all 500 participants over the course of 24 hours—a day in the life of the city. As people travel through the city, they leave traces. The white traces are where they were traveling fast, for instance by driving or taking public transportation. The blue traces are where they were traveling slowly, for instance walking, running or cycling. The resulting visualization shows the slow and fast areas in Seattle.
• #67: The second video is called Shopping for Health. It spans an entire week, and focuses on supermarkets, represented by the blue bubbles. As people get close to the supermarkets, the corresponding bubbles grow in size. By the end of the week, you can see which supermarkets were the most accessible.
• #70: The takeaway from these visualizations is that there is a high degree of mobility among the participants of the study. This challenges the concept of food deserts. The theory assumes that people are less healthy when they don’t have access to a quality supermarket in the neighborhood they live in. The study and the visualization both show that people will travel for their various needs, motivated by factors such as cost, or quality.