Internet of Things Data: From Glut to Glory

@mitrepepper – Making new friends during #quarantine

Author: Dr. Qiang Lin

According to the independent website Government Technology, “Many government agencies have an unstructured data nightmare, with bits and bytes scattered across servers, clouds and hypervisors.” Vernor Vinge calls this so-called nightmare “data glut.” In an era of Internet of Things (IoT), the data glut may not get better, and, in fact, is likely to get worse.

A study from Juniper Research found that the total number of IoT connections will reach 83 billion by 2024, rising from 35 billion connections in 2020. Given that IoT and data are intrinsically linked together, by 2025, the total data volume generated by IoT devices worldwide is forecast to reach 79.4 Zettabytes (a unit equal to one trillion Gigabytes).

Can we turn data glut to data glory? To answer this question, we need to understand the nature of IoT data and how analytics can turn raw data into information that we can actually use.

IoT Data Characteristics

IoT sensors and devices generate a huge amount of data automatically and continuously. The value of IoT data is that it is really big data. In other words, raw IoT data is combined with data from other sources, such as personal and business data, to generate new information and business intelligence. The independent website Impact summarizes these data characteristics as 7Vs:

Volume: According to Wards Auto, a single autonomous test vehicle produces about 30 Terabytes data per day – 3,000 times Twitter’s daily data. Therefore, IoT data volume can be huge, depending on the applications.
Velocity: “Even the best drivers can struggle to squeeze into tricky parking spots, but 360-degree camera systems can make low-speed scuffs a thing of the past,”, reports Alex Leanse of MOTORTREND. A 360-degree camera system consists of several video cameras strategically placed around the vehicle. For example, one is in the front of the car, one on the back of the car and two on the car’s two sideview mirrors. Then, software captures and processes the view coming from each camera and stitches them in one 3D image on your infotainment screen when the car is in reverse mode. Therefore, the speed at which IoT data is processed and becomes accessible must be timely or even real-time.
Variety and Variability: IoT often involves diverse types of data from different sensors, such as temperature readings from your home’s thermostat or a chemical plant’s steam pipeline, proximity reading for your car’s auto-braking system, pressure readings from barometers or semiconductor process equipment, optical sensing data from motion detector or LiDAR (i.e., 3D laser sensing), and so on. IoT also tends to involve multiple data sources, such as 360-degree camera systems. In addition, IoT data contains both spatial and time information. For example, the water quality sample data collected from a reservoir consists of sampling time and exact location (i.e., latitude and longitude).
Veracity: Many factors may affect IoT data quality. Examples include IoT devices and applications deployed on a global scale (like Tesla cars); lack of computing, storage and power resources; intermittent loss of connection, and sensor failure. All these factors can result in various IoT data errors, such as data readings deviating from expected values by a constant offset or drifting error, no reading, or fixed reading due to a jammed sensor or sensor crash, correct data with wrong timing, and so on.
Visualization: Large-scale IoT applications employ a large number of sensors, resulting in a very large amount of collected data. Data visualization gives us a clear idea of what the information means by displaying its visual context through maps and graphs. For example, visualization supports the monitoring of IoT devices and infrastructure for better performance on IoT data flow.
Value: Data analytics make it possible to derive value from the huge amount of IoT data—to get from raw material to usable information.

Data Analytics

According to Daniel Harris, “While new sensor, mobile and wireless technologies are driving the evolution of IoT, the true business value of the IoT lies in big data analytics rather than hardware novelties.” In other words, the data generated from IoT devices is only valuable if it actually gets analyzed, which brings data analytics into the picture.

Data analytics refers to the process of examining datasets to draw conclusions about the information they contain. Different types of data analytics can benefit IoT investments:

Streaming Analytics: Primex’s AcuRite division makes several models of Internet- connected weather stations with tons of sensors, including those for rainfall, wind speed and direction, temperature, humidity, barometric pressure, and even lightning. These sensor data are then uploaded to the cloud to calculate derived values such as dew point, wind chill, heat index, rain rate and wind speed average before streaming these updates to the user’s web and mobile devices. In 2017, Primex deployed a new streaming analytics tool – SQLstream Blaze on Amazon Web Services – which resulted in better performance with significant cost savings.
Spatial and Time Series Analytics: Lack of free parking spaces is one of the major reasons for traffic jams in big cities. “It is estimated that about 30 percent of the cars circling a city at any given time are doing so as drivers look for parking.” Innovative smart parking technology combined with IoT connectivity helps solve this problem. Installed IoT sensors determine where empty parking spaces are located. Then, the empty parking spots data is transmitted over a wireless connection to a cloud server. All the data from the parking lot is collected and analyzed in real-time to produce a map of available spaces made available to drivers looking for a space.
Predictive Analytics: In a factory setting, downtime can be costly, and machine failure can be expensive and even dangerous to workers. IoT sensors attached to factory robots can track variables, such as vibration, temperature, and machine timing, and feed that data into a predictive analytics platform. Then, the data is analyzed to predict when a particular machine will need maintenance so that factory managers can catch a device before it fails – reducing the risk of downtime or a more expensive repair.
Prescriptive Analytics: Upsolver describes how prescriptive analytics suggest actions based on the result of a prediction or diagnosis about how to optimize or fix something:
- This machine is 80% likely to fail in next 12 hours. How should I prevent this?
- This machine is creating too many defective components. How can I avoid this?
- This design is resulting in too many manufacturing issues. How can I improve it?
Artificial Intelligence (AI): Autonomous driving is an excellent example of how AI and IoT can work in concert. Autonomous vehicles are loaded with sensors necessary to collect vast amounts of data about their surroundings constantly. This data is processed for intelligent insights using AI models that enable the vehicle’s navigation system to negotiate surroundings and choose the optimal path in real-time.

Correct use of data analytics will help Federal agencies and enterprises turn data glut to data glory. As 2021.AI’s Peter Sondergaard reminds us, “Information is the oil of the 21st century, and analytics is the combustion engine.”

MITRE has worked for more than a decade to show how masses of information, combined with sophisticated analytics, can provide our sponsors with powerful decision-making tools. For example, we’ve used the data from a remote-sensing satellite, such as the data on water quality, temperature, and bacteria to help researchers dramatically improve monitoring of the Chesapeake Bay coastline. In addition, MITRE’s Data and Human-Centered Solutions Lab applies and extends the state of the art in information management, analytics, behavioral sciences, and more – working with sponsors and turning information into value.

Qiang Lin is an Information Systems Engineer at MITRE and Adjunct Faculty in the Department of Electrical and Computer Engineering at George Mason University. He worked at Deloitte and SAIC previously, and has a PhD in Computer Engineering from West Virginia University.

MITRE’s mission-driven team is dedicated to solving problems for a safer world. Learn more about MITRE.

Internet of Things Security: Challenges and Solutions

5G: How it Works and What it Brings

The Power of Geospatial Data In Developing Countries

Dr. David E. Willmes on Solving Global Food Insecurity

Will Quantum Computers Revolutionize My Daily Life? Not in the Ways You Might Think

The Person at the Other End of the Data

A Social Scientist Examines the Role of Technology in our Lives

Building Smarter Machines by Getting Smarter About the Brain

Marcie Zaharee and MITRE’s Open Innovation Challenge

Another Update on the Aviation Industry with Michael Wells and Bob Brents

Interview with Dan Ward, Debra Zides, and Lorna Tedder on streamlining acquisitions