Data analytics at the edge: benefits and challenges

By Data Science Salon

Data analytics at the edge is the trend of connecting autonomous vehicles, wearable devices and automated oil rigs that are operating with no human assistance. The benefits are huge, yet so are the challenges. 

According to the Ericsson Mobility Report 2021, mobile networks today carry nearly 300 times more mobile data traffic than in 2011. The IDC predicts that by 2025 there will be 55.7 billion connected devices in the world with 75% of which being connected to an IoT platform. The data generated by connected IoT devices is estimated to reach 73.1 zettabytes by 2025.

This fuels the rising interest in data analytics at the edge. 

What is data analytics at the edge?

Edge computing refers to performing operations closer to the device that is gathering the data - be that a digital camera, the IoT smart device, a smartphone or an autonomous vehicle. The “edge” is neither about the size nor about the processing power of the device but the distance from the core data center of the company. 

Data analytics at the edge is basically about enabling the edge device to analyze the data and process it before sending it further. The analysis needs to be done in a fully automated way, either using a system of preprogrammed triggers, heuristics or machine learning-based solutions. 

With the increasing number of connected devices and data flowing through the network, moving the analytics closer to the edge comes with numerous advantages for the company. 

Benefits of data analytics at the edge

Data analytics at the edge device reduces the need to transfer the data to and from the data center in the cloud. This can be a game-changer considering the effects listed below. 

Reducing the costs

While data transfer is getting cheaper, it is not free. Also, the more remote the location, the internet connection is less reliable. This results in the rising costs of data transfer. 

Also, data needs to be processed and stored, further building up the costs. Data analytics at the edge comes as a great way to reduce spending. 

A good example comes from Axis Communications Zipstream technology used in surveillance cameras. The data from CCTV is usually about hours of nothing occuring and seconds of important events like a burglary, a fire or any other incident. 

The Zipstream technology is an edge analytics tool that cuts unnecessary information and keeps what’s important in high quality. The company claims that the technology reduces the costs of bandwidth and storage by up to 50%. 

Reducing the latency

The latency occurs when the time required to transfer the data impacts the operations. In numerous cases reducing it can be not only an improvement of the business efficiency but a life-saver. 

A good example of the edge analytics-powered latency reduction comes from Trimble Transportation, the company that delivers the solutions for monitoring the cargo fleet. If any sensor mounted in the truck shows a dangerous situation - for example: a sudden tire pressure reduction - every second matters. The latency reduction delivered by the edge analytics can be a real-life-saver, both for the truck driver or other people on the road. 

Analytics in the remote locations

A great example of edge computing leveraged to optimize the operations in an isolated localization is the oil rig. The installation needs to have enough computing power to analyze the data flowing from the sensors and notify the operators only when detecting the anomaly. 

The pinnacle of this approach comes as a fully automated oil platform Oseberg H operated by Norwegian giant Equinor (formerly Statoil), completely unmanned and controlled remotely. 

Challenges in analytics on the edge

While delivering the benefits mentioned above, edge analytics is not a panacea and comes with challenges not to be overlooked. 


One of the key challenges with the analytics at the edge is the reduced access to the raw data gathered from sensors. This can be good enough in many cases. On the other hand, training the machine learning algorithms (ex. aimed at predictive maintenance) may be nearly impossible without access to the raw sensor data. 

The machine needs to be heavily tested to ensure the accuracy of the delivered analysis. And there may not be enough time or resources to do so. 


Analytics can be easy when having access to the data center and powerful computers. In the edge device, the hardware is a limiting factor, forcing the development teams to deliver optimized solutions, consuming less power, and using limited storage that is available on the device.


Last but not least, edge devices are a frequent target for cybercriminals. The cybersecurity company Kaspersky detected 1.5 billion attacks on IoT devices. The goal is either to hijack the computing power or to get access to the company’s data and resources. 

Empowering the IoT with analytics capabilities makes them more vulnerable for the attack delivering more computing power and resources when compromised. 


Edge analytics comes with multiple advantages such as reducing the costs and latency. Yet, it needs to be done in a secure, reliable and tested way to provide all the benefits and mitigate the risks. 

Get the latest data science news and resources every Friday right to your inbox!