Burgers, IoT and cloud computing for the perfect restaurant experience

Peter Pane wants its guests to feel at home in its restaurants, which is the reason why we addressed two central questions in the project: How can we create added value for Peter Pane and its customers? What data is best for measuring a feel-good atmosphere? We therefore started by collecting data on obvious feel-good factors.

• Temperature – Is the restaurant too cold or too hot?

• Humidity – Is it too stuffy or does dry air from the heating irritate people’s airways?

• CO₂ content – Is the air fresh or stale and unhealthy?

• Volume – Is the noise level too high or can you talk comfortably?

And what about smell? Interesting sensors are gradually coming onto the market that can be used, for example, to detect unpleasant odors. However, since we aimed to minimize technical risks and were unable to assess how far we had already come with the basic feel-good factors at the start of the project, we first decided not to use such sensors and will re-evaluate them after the project launches.

Peter Pane aims to ensure long-term customer satisfaction. A pleasant atmosphere is part of this. Simple measures can be immediately identified through continuous measurement. This allows the heating or music volume to be adjusted or fresh air to be provided when the relevant thresholds are reached. At the same time, we also set out to find out how the parameters affect how guests order – especially when the atmosphere is louder, stuffier, warmer or colder than usual. That's why it was not enough to use conventional devices. Instead, the data had to be continuously collected and stored in a database.

Collecting measuring points with sensors, processing them via the Internet of Things, and developing new insights and recommendations for action from this data requires a cross-functional team covering a wide range of disciplines. codecentric not only has suitable experts for such a project, but also contributed a great deal of experience from the IoT field to successfully implement the project.

To this end, we equipped Raspberry Pis with the appropriate sensors, wrote the necessary software and printed suitable, inconspicuous housings for our sensor configuration using 3D printers. We first implemented a suitable device management system to allow remote maintenance of the Pis. We then set up the cloud infrastructure in the IONOS cloud, where all sensor data is sent via HiveMQ to a Postgres database and then visualized in Superset dashboards. We used both Kubernetes and Stackable to process large amounts of data.

With this setup, we installed the four sensor packages inconspicuously above the ceiling paneling in the Cologne branch of Peter Pane. The locations of the Raspberry Pis were chosen to ensure good coverage of the restaurant and also to reveal differences between different areas. We only had a two-hour window to set it up. Setup went smoothly, so that all sensor data could go online on time.