With this project we are tackling the challenge for biodiversity and sustainability from the Smithsonian Institution.
The inspiration for this project was based on an initiative in Switzerland which is focused on building biodiversity zones in urban environments. Their setups include different installations to enable biodiversity like breeding boxes for birds, nesting places for bats, wild bushes for bees and many more. Unluckily for the initiator, these are basically "fire and forget" measures as he doesn't get any insights on the usage of the installation and their effectiveness.
Our projects main goal is to capture data around these installations, store the data and provide actionable insights to the experts, so they can improve the installation, learn about the behavior of the animals living in these zones and use the insights for future projects. Benjamin Kämpfen is an environmental scientist and the co-lead of a local BirdLife Group in Zurich, he joined the project to provide the necessary biodiversity expertise and helped us to define the use case for this project. He will also be the expert analyzing and qualifying the data that the sensors are gathering.
This picture shows an example of such a concept for a biodiversity area in an urban environment that Benjamin and his team creates for his customers.
What it does
Our solution captures activity, temperature and humidity with a sensor device which is mounted in a nesting box for birds. For the Hackathon we equipped two nesting boxes with sensors first of all to gather the data for Benjamin, but to also evaluate the data quality and the durability of the hardware.
The results of this hackathon are:
- A database with the processed and enriched sensor data for long term data storage and easy access for everybody who would like to analyze and work with the data.
- A visualization of the data based on PowerBI. The goal is to continuously improve the dashboards with the feedback of Benjamin and other environmental experts.
With the gathered data Benjamin is looking to gain new insights into the birds behavior, like the timeframe when newly born birds hatch or the moment when they leave the nesting box for the first time. He hopes to use these findings to improve efficiency of the biodiversity measures in Zurich and also make the installations more tangible to the public.
How we built it
At first wanted to build our own devices based on a Raspberry Pi and external sensors but with the limited time of the hackathon and the challenge of the power supply, we decided to drop the idea and go for "off-the-shelf" hardware. Luckily we found a local vendor who managed to supply us with the necessary hardware within hours. The devices with built-in sensors are mounted in the nesting boxes and are battery powered.
The IoT devices are connected via LoRaWAN with a centralized gateway, which is mounted in the vicinity of the sensors and is connected via 4G to the internet. LoRaWAN is based on 868 MHz radio band and the sensors are only sending data over radio every 15 minutes for a very short timeframe to make sure that there is no impact on the local wildlife because of the radio waves. This communication technology is also very power efficient, so our plan is to leave the sensors in the boxes for at least 12 month without the need of maintenance or battery changes.
The gateway translates the LoRaWAN protocol into the AMQP protocol to forward it to the Azure IoT Hub in the Microsoft Azure Cloud. The IoT Hub is the central ingestion point for IoT data in Azure, from there we can very easily integrate the data with other components. In our case we had to use Azure Stream Analytics to do some transformations on the sensor data, because the data is sent as a message encoded in hex from the sensor device and we had to translate it into decimal values. The Stream Analytics component also makes sure that the message is put into the right format to be directly saved into a SQL database for long time data storage. In the database we implemented a data model, which allows us to join meta data with the sensor information to put it into the right context.
To visualize the data we leveraged PowerBI, the business intelligence application from Microsoft, because it was the easiest way for us to generate dashboards on the gathered data and share the results with Benjamin on the stakeholder side. PowerBI directly pulls the data and the data model from the SQL database, so no further processing was necessary for the data visualization part. We also wanted to give Benjamin the opportunity to analyze and visualize the data in a self-service approach.
The picture above shows a first draft of such a dashboard that we created based on the incoming data. Together with Benjamin we will continuously improve the dashboard to generate insights based on his domain know-how.
Challenges we ran into
Of the Shelf Hardware
We were first planning on building our own hardware based on Raspberry Pi and external sensors but after the first day we figured out, that this is way to complicated for the given timeframe and especially the power supply would have been way to complicated, as there is no feasible battery powered solution available. We did some research and managed and to find a vendor who supplied us with industrialized "off the shelf" devices which are capable of measuring the necessary sensor values for us.
Listen to the domain experts
Before the start of the hackathon we brainstormed a lot about the use cases we could implement in such a biodiversity zone and the sensors we could use, but as we are no experts in this field, we weren't too successful. Luckily we managed to get Benjamin on the team and he supported us in identifying good use cases for the implementation.
Tech meets Biodiversity
When tech people meet experts for biodiversity and sustainability both parties first need to find a common understanding of a topic, so that everybody aims to achieve the same goal.
Accomplishments that we're proud of
At the end of the hackathon we managed to fully equip two birds nesting boxes with sensor devices, capture the data and forward it via the LoRaWAN communication protocol to the Azure cloud, where we compute, store and visualize the data. This setup is actively capturing data and Benjamin can start working with it. We are extremely proud that we managed to implement this solution in the given time frame, even though we know there is much more work ahead of us.
This Hackathon didn't just help us to build a great foundation on the technical side, it also brought together engineers and biodiversity experts to work for one common goal and helped us build a great partnership. We are sure that this was just the beginning of something big, that we are going to achieve together.
What we learned
Combining the internet of things with biodiversity and sustainability is a greenfield. There is very little experience around and it was great for us to explore this field together with Benjamin, our domain expert. And this collaboration between an environmentalist and tech-guys was the greatest learning for us. As a single party none of us would have been successful, but together me managed to achieve something great.
What's next for IoT for Biodiversity and Sustainability
The sensors in the birds nesting box are just the beginning of this journey, we are already planning to use similar sensors in nesting boxes for bats and bee hives. With LoRaWAN as communication technology there are countless other sensors types available, we just need to identify the use cases.
We would also like to start testing camera and microphone based solutions with AI and Machine Learning on the edge to help us identify which species is breeding in the nesting boxes to allow the experts even deeper insights into the behavior of the animals. Its also important for us to install as many sensors as possible to build a large database, which allows a holistic view on the biodiversity measures and the animals living in it, to generate actionable insights.
There are countless possibilities for the use of sensors in this field and we will work closely with the environmental experts to explore the use cases and rapidly implement proof of concepts.
Remark on Demo
We tried to publish our demo dashboard to the public, unluckily our PowerBI license didn't allow us to do so. At the moment we can only share it with people with a PowerBI account. Please reach out to us if you are interested to get access. We are more than happy to share the dashboard.