Report IoT sensor analytics workshop

 

Recently a customer of SMT organized an IT DevDay event. The day had a packed program with several hands-on labs, keynotes and presentations. SMT presented a workshop about IoT sensors and showed its practical application by collecting data from the building of the customer. The IoT sensors generate data from various devices which can help to understand their performance and for instance predict maintenance. These insights are great assets for analyzing the business, how to increase efficiency or in which way organizations can benefit from it.

The sensors were provided by Anything Connected. They created an internet connected sensor sticker which measures several different variables such as electric activity. It wirelessly measures what electric devices are doing, so no need to pause the machine and connect something into the circuit. The collected data can be delivered to any data platform for analytics. As the data is collected on the outside of the device, no personal or privacy sensitive information is involved. In short, they are non-invasive sensors for monitoring external activity.

Sensors have been placed at various locations in the office building: the elevator, the coffee machine and the turnstiles to enter the building. The sensors were easily attached and removed to any object with industrial tape. They measure acceleration, vibration, position, electrical activity and temperature. This data was collected and loaded into the SMT Enriched Data Analytics Platform® for which Splunk is used.

This is a data analytics platform which brings structure to machine data from any application, server or network. It gives the possibility to process unstructured machine data and make it insightful. This data can be used for analysis, real-time insights and predictions.

The benefit of using the Splunk platform in this situation is that it enables fast exploration and analysis of the data. Plus, it is scalable; the same technology can be used in both the proof of concept as well as in a scaled-up production environment. A machine learning algorithm can easily be applied from the available Machine Learning Toolkit.

Around lunch time the sensors were put on the turnstiles, for which access passes are used. As the turnstiles turn one way when people go out and the other way if people come in, it was possible to measure the number of people going out and coming back in. In this case, the number of people going out for lunch also returned to their desk in the afternoon. No one was missing.

Through the sensors on the elevator it was possible to measure the number of times the doors opened and closed. If the elevator moved from one floor to the next or all the way from level one to the fifth floor. The number of trips and how much time it is used is highly interesting information for the maintenance department.

Within a short period of time, it has been demonstrated that data is everywhere and that this can be quickly transformed into information that is of value at various places in the organization. The total movement time of the elevator is, for example, an interesting metric in negotiations with an elevator service organization. A change in the trend of using the elevator is a factor that can be used to measure the effectiveness of a behavioral campaign (“Take the stairs!”).

The sensors provide information for preventive maintenance, but there is also a lot of additional information that can be obtained from the same data set. It can also measure behavior and the usage of any device. The SMT-consultants have years of experience using Splunk, so they were quickly able to collect the data, create a clear dashboard and retrieve some valuable insights. Hence, it is possible to retrieve valuable information from the collected data about Information Security, Business Analytics or IT Operations.

A concrete example of the analysis possibilities with IoT sensor data was given in a short time. Raw measurement values are quickly converted into useful and valuable information. For this organization, which itself often works with sensors, it was preferable to demonstrate the potential value of sensor data in a quick and playful way. Because of the techniques used, the sensor types and the SMT EDAP®, SMT was able to provide these first insights within two days. The customer gained new insights into the usage and application of IoT sensors. Moreover, they have experienced how the right tools can support them to get results quickly.

 

May 2019