Cloud Computing
Electronic devices equipped with numerous sensors are omnipresent today. Countless amounts of data are generated and collected every day, in our private lives, in professional sport and in everyday business life. But what is behind this data, what is it used for? And how can the collected data be optimally utilised?
The term "big data" has become an integral part of the vocabulary of modern companies. Data is collected according to the principle of "more is more". However, collecting and storing data is easier said than done. It is often associated with high costs and additional effort. Once the data is finally collected, the question arises as to what to do with it. It is therefore important to collect the data that tells a story. Big data analytics is similar to telling a story. A lot of data can make an excellent story with a lot of substance. Based on this, fundamental management decisions can be made or failures and response times can be minimised. On the other hand, misinterpreted data tells a story that is not understood and raises more questions than answers.
In spring 2018, we at Swisscom Analytics created a proof of concept (PoC) for Swisscom's property company. The aim was to equip the meeting rooms within Swisscom with certain sensors in order to create an optimal working environment and achieve efficient energy utilisation.
Data sources
Before the PoC could start, it was necessary to clarify which sensors would be useful for the feasibility test. We carried out the PoC with the following sensors:
The temperature sensor should be used to check whether the temperature is comfortable or whether it should be adjusted upwards or downwards. The humidity sensor does the same for the air humidity. The light sensor can be used to check whether the lighting in the room is on or off. The ultrasonic sensor could be used to check whether someone was entering or leaving the room.
One requirement of the PoC was to use Hortonworks Hadoop. Within these specifications, we were free to choose the frameworks. All frameworks used are included in the standard installation of Hortonworks Hadoop.
Kafka
Kafka was used to load the data from the sensors into the Hadoop environment. Kafka is a distributed streaming platform that is ideal for real-time streaming applications. Kafka is also compact and fault-tolerant.
Flume
In order for the data from Kafka to be available within the Hadoop file system, it must be loaded from the so-called Kafka topic into the HDFS. Flume is used for this.
Hive
As access from the visualisation frameworks is easier if the queries can run via SQL queries, Hive is used. Hive allows the data in HDFS to be queried using the familiar SQL syntax.
Visualisierungsframework
The visualisation framework that the user feels most comfortable with can be used, but care should be taken to ensure that the framework is compatible with Hive. Zeppelin was used for the PoC. Other options are SAP Lumira Discovery, Tableau, etc.
The current sensors provide information on whether it is too hot or too cold, whether the humidity is at a comfortable level for users and how much light is coming in.
The ultrasonic sensor is more exciting. This can be used to check whether the door is open or closed or how many people have travelled through the room. This would require counting how often the sound barrier has been broken. The ultrasonic sensors could also be positioned above the seats so that it is possible to analyse where a person is sitting.
For example, the data obtained can be used to specifically deploy humidifiers if the humidity is too low. The light can also be dimmed when the sun is shining more strongly, which can also save electricity costs. If the ultrasonic sensor detects whether there are still people in the meeting room, the light can be switched on or off accordingly.
In addition to the potential offered by the existing sensors, there are a number of other sensors that could be installed. For example, noise sensors to measure the noise level.
Swisscom Analytics will be happy to prepare a feasibility study for your big data requirements and support you on the way to successfully analysing your collected data.
The contact persons at Swisscom are as follows:
Tim Giger (tim.giger@swisscom.com(opens in new tab)) – Hadoop Lead
Matthias Mohler (matthias.mohler@swisscom.com(opens in new tab)) – Analytics Lead
Lukas Heusser (lukas.heusser@swisscom.com(opens in new tab)) – Analytics Consultant
Analytics Consultant
Find the job or career to suit you. A career where you can make a difference and continue your personal development.
What you do is who we are.
