Any discussion of analytics today has to include some consideration of the impact of the industrial Internet of...
Things, or IIoT, the proliferation of sensors and connected devices that are the source of much of the big data that analytics feeds on. What's different -- or not different -- about IoT sensors and devices? Where and how are they being used? And how does this all fit with big data and analytics?
How IoT sensors differ
The two characteristics of the new breed of sensors that are driving the revolution in IIoT, big data and analytics are that they are inexpensive, and they readily connect to and through the Internet. While many sensors and devices -- programmable logic controllers, for example -- have had either Internet or Ethernet LAN connectivity for years, connectivity has now become ubiquitous -- and cheap.
And that makes all the difference. Whereas a few years ago it might have cost thousands to wire up a small section of a plant, or a piece of equipment to gather a few basic pieces of data, today they can be loaded with measurement devices, counters, video with image recognition software and more, for a small fraction of the previous cost. Everything communicates "plug and play" through IP connections. The biggest challenge with IoT sensors is in making effective use of the data.
IoT devices add context, automatic response
Ubiquitous and inexpensive connectivity extends beyond simple sensors. Lots of other kinds of devices, all connected and inexpensive, add to the basic data and make it more meaningful by adding context -- GPS location, for example -- thereby going beyond straightforward data collection. Actuators can accept signals from the network and actually make things happen with switches, solenoids, equipment adjustments and more sophisticated transducers converting digital signals into automatic responses to conditions identified by software operating on sensor data feeds. This in itself is not really new either, but wider availability of more types of devices for far less cost and far easier implementation change the playing field. A relatively high level of automation is now available to small and midsize organizations that don't have a cadre of technical experts on hand.
The devices category also includes tablets and smartphones to access, analyze and display the data among a much wider universe of people, anywhere, anytime.
Where are they?
The IoT sensors and devices are replacing manual reporting, and absence of reporting, in obvious places where visibility into operations can enhance performance. In the plant, IoT sensors directly attached to equipment and workstations can continuously monitor equipment cycles, timing, workpiece dimensions and other physical parameters, and more. But that's just the beginning. It is more feasible to track things outside of the plant.
Automated identification technology can record the exact contents of each box, case, pallet, container and truckload along with serial numbers; location within the truck or container; ambient conditions, such as temperature, humidity; and location through GPS. A truck halfway from New York to California could be rerouted to Dallas to accommodate a customer emergency and the exact goods delivered, with customer service having full knowledge of exactly what happened and how to coordinate resupply of the original intended recipient.
Let's say that a company has repeatedly experienced shipping damage on a delicate piece of equipment despite what they thought was good packaging and loading into the truck. It can equip the next shipment with accelerometers inside and outside the package, video cameras, pressure sensors and more, combined with GPS and precision timepieces to monitor the entire journey in great detail to identify when and how the damage occurs.
Or imagine a machine that is experiencing problems -- a molding machine that is producing damaged parts. Maybe the best way, or only way, to find out what is really going on is to observe the internal workings of the molded-part removal process mechanism. An inexpensive "lipstick camera" -- the size and shape of a tube of lipstick, or smaller -- might be mounted inside the machine and the resulting video can be studied in slow motion or one frame at a time to identify the cause of the damage.
Bringing it all together
The above examples may not relate to analytics in the traditional sense, but serve to illustrate the extent to which new sensor and device technologies extend visibility to places where it was not possible or affordable before. You'll have to use your imagination to envision how detailed, real-time, and location or time-logged data can provide a new and unique window into processes, activities, goods and facilities, and how that might open the possibilities for better management, earlier and more informed preventive action, and opportunities to improve performance.
The masses of new data, and the variety of new data types, including video, click-streams and unstructured text, strain the capabilities of yesterday's data-management and analysis capabilities. Thus, the new breed of analytics is evolving to not only "manage" big data, but also to bring about new ways to combine traditional data with big data to generate useful business information.
To make it usable, modern analytics must be "consumerized," so that everyday users can build their own dashboards and analyses without the help of programmers or data scientists. Data visualization -- turning millions or billions of data points into something that a mere human can interpret and gain intelligence from -- is a core task for big data analytics.
See a guide to machine-to-machine technology
Read more by this author on mobile logistics