Measuring and Improving Health with Sleep Data

View all blog posts under Business | View all blog posts under Infographics

An estimated 35% of adults do not get enough sleep, according to the Centers for Disease Control and Prevention. Teenagers have even worse sleep habits, increasing their likelihood of poor sleep later in life. Individuals of all ages need adequate sleep every night to decrease their risk of mental illness and other physical conditions. Researchers and tech companies are harnessing the power of sleep data to identify the causes of poor sleep quality and help individuals get the sleep they need.

To learn more, check out the infographic below created by Villanova University’s Master of Science in Analytics program.

How data is being used to improve individual sleep.

Add This Infographic to Your Site

<p style="clear:both;margin-bottom:20px;"><a href="" rel="noreferrer" target="_blank"><img src="" alt="How data is being used to improve individual sleep." style="max-width:100%;" /></a></p><p style="clear:both;margin-bottom:20px;"><a href="" rel="noreferrer" target="_blank">Villanova University </a></p>

Sleep Data 101

A variety of devices can collect and analyze sleep data, in both clinical and home settings.

Data Collected While You Sleep

Sleep data can be collected through a polysomnography (PSG), which uses sensors to collect data on metrics like brain activity, airflow, body position, heart rate, and eye movement. Another device, a videosomnography (VSG), uses a range of video-based methods to record and score sleep behaviors. Other devices used can collect data on movement patterns, heart rate, and air pressure to track sleep metrics.

How Data is Collected, Analyzed and Applied

Data is commonly gathered via two methods. The first is clinical methods, which features elements like PSG, VSG, and continuous positive air pressure (CPAP). The second is customer methods, such as diaries, accelerometry to track movement patterns, and heart rate.

There are also several forms of data storage and curation. These include transferring extracted data from source systems to data warehouses, and stripping or encrypting personal or identifying information from sensitive data. Another form of data storage involves data selection, which can include filtering, aggregation, and querying. Additionally, sleep data can be processed via different methodologies. These include normalization, transformation, integration, validation, and featurization.

Data modeling is also a key part of the sleep data process, and it’s approached in three distinct ways. One way is the heuristic approach, an approximation method used to develop ideal solutions when presented with different kinds of problems. The second approach is machine/statistical learning, which is the process of using statistics to find patterns in large datasets. Another modeling approach is deep/transfer learning, a subset of machine learning that uses artificial intelligence (AI) in processing data and creating patterns to inform decision-making.

Once the data has been collected and analyzed, it can be applied through medical channels such as clinical support systems, research purposes, or interventions. It can also be applied via consumer means in the form of recommendation systems and lifestyle products.

Tracking Sleep Data

Though technology is often blamed for causing sleep problems, wearable trackers and other devices can help people track sleep phases and offer insights to improve their sleep quality.

How Sleep Trackers Work

Common devices monitor several important metrics. For instance, they track sleep duration by recording user activity from when they fall asleep to when they wake up. They also measure sleep quality to detect interrupted sleep and track when the user is restless or waking up during the night. Additionally, sleep trackers track sleep phases to time the user’s alarm to go off during a period of less deep sleep. They also can account for various environmental factors such as light and temperature levels in the bedroom. Finally, they account for lifestyle factors such as stress levels, mealtimes, and caffeine consumption.

There are also several emerging sleep-sensing technologies poised to further the concept of sleep monitoring. These include ultrasound sensors, smartwatches, bed sensors, wireless EEG devices, mobile health devices, and RF sensors.

Waking Up To the Future of Data In Sleep Research

Unfortunately, sleep trackers can’t measure sleep directly. They can only measure inactivity to estimate sleep. To advance sleep science, scientists are using big data and offering deeper insights to help individuals with sleep problems.

Insights Unveiled by Sleep Data

A study of over 10,000 people ages 18-100 from around the world required participants to report how many hours they slept the night before and complete a series of tests. The surveys ultimately indicated that sleeping seven to eight hours in an optimal sleep level, and sleeping too little or too much can have a negative cognitive impact.

Additionally, Fitbit analyzed data on 6 billion nights of its American users’ sleep to determine that men sleep slightly less than women, younger people typically get more deep sleep nightly than older people, and inconsistent sleep patterns negatively affect sleep.

5 Ways Big Data is Being Used to Improve Sleep

One of the key ways sleep can be improved through big data is by improving knowledge and treatment of sleep conditions. Another way is by identifying root causes of sleep disorders. A third means of improvement is the linking of behaviors to sleep quality. Big data is also used to improve mattress design, which can help improve sleep. Finally, big data allows for the personalization of content and recommendations for better sleep.

In addition, big data has several potential applications. These include predicting risk, targeting sleep interventions, and performing sleep disorder surveillance.

Information That Helps to Rest Easy

The intersection of big data and the field of sleep study promises to offer patients more accurate diagnoses and treatments. This should bring some comfort to those ready for a good night’s sleep.

CNBC, What You Should Know About Getting a Good Night’s Sleep, From a Company That Has Analyzed 1 Million Nights of Sleep Data

IBM, Data Aggregation

Import, What is Data Normalization and Why Is It Important?

Innovation Enterprise Channels, The Big Sleep: Big Data Scientists Tackle Lack of Quality Shut Eye

Investopedia, Data Anonymization

Investopedia, Deep Learning

Johns Hopkins Medicine, Do Sleep Trackers Really Work?

Microsoft Support, Create a Simple Select Query

MIT Technology Review, What is Machine Learning?

Nature, The Future of Sleep Health: A Data-Driven Revolution in Sleep Science and Medicine

NCBI, SleepOMICS: How Big Data Can Revolutionize Sleep Science

New Gen Apps, Data Science Modeling and Importance in Featurization

Oracle, Database Data Warehousing Guide

Sleep Advisor, 54 Shocking Sleep Statistics and Trends for 2020

The Sleep Doctor, 5 Surprising Takeaways From The ‘World’s Largest’ Sleep Study

Stitch, What is Data Transformation: Definition, Benefits, and Uses

Talend, What is Data Integration?

Techopedia, Data Filtering

Wiley Online Library, The Heuristic Approach and Why We Use It

World Economic forum, Fitbit Analyzed Data on 6 Billion Nights of Sleep – with Fascinating Results