Sensors are everywhere these days. How do you work with a variety of raw sensor data formats?
What are sensors?
Sensors are everywhere these days – cars, smartphones, gas stations, televisions, manufacturing plants, solar panels, water treatment facilities, hospitals, and many more. Everywhere!
Sensors come in many shapes and sizes measuring lots of parameters including temperature, vibration, humidity, location, pressure, motion, and ambient light. They can also be wired or wireless and range in size from very large for hardened industrial applications to tiny for embedded applications like smart phones.
Sensors output a variety of data formats
With that wide variety of sensors comes an equally wide variety of data and data formats emanating from them. Most sensors output a number or array of numbers that can be fed into an application for processing, display, alarms, or analysis. However, this number can represent a variety of different values and may have to be scaled or converted (Transformed from the Extract, Transform, Load, or ETL model) before passing to another application. For example, the output from a temperature sensor could be in degrees, 1/10 degrees, Fahrenheit, Celsius, etc. Sensor manufacturers often offer software to convert their data to common formats, but if you use sensors from a variety of companies chances are you’ll have to do data transformation before processing.
How do you transform sensor data?
In addition to transforming sensor data, it often requires some computation before it can be used by another application. For example, vibration or current data may be averaged over a time window, or converted from time-domain data to frequency domain data, etc. Computation is inherent to working with sensor data, and most standard ETL tools are not adequate. In fact, most ETL tools are designed to work with transactional data that require extraction and transformation, but not computation.
How do you deal with all these different types or sensor data? Software developers tackle the problem by writing a dedicated algorithm in a programming language. This approach solves the problem but requires programming expertise, is inflexible, and difficult to maintain. Engineers might take a different approach, using a variety of low-code or no-code languages to convert the data. The advantages here are faster development time, far less programming expertise required, and easier maintenance when adding sensors or modifying the system design.
Use low-code to work with raw sensor data
Enter Node-RED, an open source, low-code design framework designed at IBM nearly 10 years ago. With a community of well over 10,000 users, over 2 million downloads, and more than 3,000 contributed nodes, Node-RED is the fastest growing low-code system in the world. Not only is Node-RED easier to use than full-code approaches, with so many nodes available, chances are high that you can find one to do any data conversion imaginable. You can even find nodes to convert raw data and transmit it through edge gateways using a variety of protocols including TCP IP, BACnet, Ethernet/IP and many others.
Prescient Designer, Prescient’s flagship product, is based on Node-RED and allows engineers and developers to acquire, extract, transform, analyze, compute, and visualize data from any source, including sensors. An enterprise grade, cloud-based platform, Prescient Designer takes the power of Node-RED and adds security, co-development, edge/cloud computing transparency, and the strongest hardware support in the industry (edge controllers, gateways, PC’s, and GPU’s). Finally, Prescient Designer has a variety of application templates, including one for data acquisition which can help you start gathering, displaying, and analyzing sensor data in minutes.
Prescient Designer – the perfect low-code design platform for sensor applications! Sign up to get a free live demo and Prescient Designer account today.