London's Natural History Museum puts on display of urban nature with AWS

As the most visited museum in the UK, the Natural History Museum in London is well-known for its iconic building housing a vast range of exhibits - 80 million items and counting, spanning billions of years. Its five million-plus annual visitors are treated to an array of striking specimens, from Hope the blue whale to an animated, roaring T-rex, a dodo that is certainly dead to Superman-conquering kryptonite. 

The museum is now turning its visitors’ attention to the natural world much closer to home with the Urban Nature Project. The project will see the museum's five-acre site in London transformed into a biologically diverse green space, along with the development of new scientific tools to monitor and protect urban nature as it responds to changing conditions.

To fulfil these objectives requires the use of new technology to rapidly gather data on biodiversity in the environment at scale. But while the museum’s team of scientists are comfortable with the science element, the technology side is less familiar. 

To help the museum build the new tools and understand the tech, it has teamed up with Amazon Web Services (AWS) as part of a five-year partnership. Dr John Tweddle, Head of the Angela Marmont Centre for UK Biodiversity at the Natural History Museum, explains:

We were looking for a tech partner who could help us bring our dream of a research ecosystem where we can stream different data types, analyze them collaboratively and share with our partners through the cloud. Chatting to AWS, straightaway they understood what we're trying to achieve for the environment. But they could also say, 'The tech you need is this and why don't you go in that direction instead?'.

Working with AWS has been an eye opener for the museum, as the process is very different to its normal approach. Tweddle says:

The idea of starting at the end with, 'What's the story you want to show?', then working backwards to develop the timelines and to deeply interrogate what your actual needs are. We’re used to starting at the beginning.

Tweddle’s 14-strong team spent quite a lot of time working with AWS staff and its delivery partner, Thoughtworks, to interrogate what the museum wanted from the system as scientists and educators. He adds:

“We're trying to get it to do a lot of different things for a lot of different purposes. The questioning approach we took, it took a good six months but at the end of that process, we had a much clearer idea on what it was we were asking. We thought we knew at the start, but we knew at the end.” 

Platform

Less than a year into the project, the organisations have developed a new data platform, built on the AWS cloud. Tweddle explains: 

We have all the building blocks, and the capability to stream different kinds of environmental data, from IoT sensors, from mobile phones to bring in visual observations of wildlife, environmental DNA samples, hundreds of thousands of observations. It's been quite a rapid-fire build of the minimum viable product. We've crammed all that into less than a year, it’s been quite intensive.

The data system is a mixture of non-relational and relational databases with various APIs. Staff at the museum – there are currently 350 scientists and 900 in all - are now being trained in how to use the AWS system, to bring in the extra functionality needed going forward as new techniques and new ways of recording nature appear, so those data streams can be plugged in as well. 

Since the system went live in May, data is now being entered for the first time via a community science project called Nature Overheard. Anyone can take part in the project, which is exploring the impact of road noise on wildlife: take a five-minute audio recording in your road, note down the location information, observe any insects and upload the data. Tweddle says: 

We can start to look at the correlations between the soundscape and insects. It's only recently we've had the tech to be able to do this, and that's the first data set going in, our first trial.

The museum is also uploading close to a million environmental DNA observations from urban settings across the UK, as part of its work to explore the hidden biodiversity in these environments to understand nature and wildlife that is present, and how and why it's changing.

The Urban Nature Project is focused on urban areas across the UK as that's where people’s connection with nature is most limited; it's also where the museum feels it can make a huge amount of nature recovery difference very quickly. 

The museum is already investigating how this infrastructure can cover all of its community science work, providing a single solution for citizen science going forward. Tweddle explains:

It's how we work with so many audiences around developing science skills and knowledge across the UK and globally. Already we are looking at how the platform can be used for a whole new science area in the museum, which is focused on UK nature recovery. We're going to get it right for the Urban Nature Project first, and then extend.”

Part of the rationale behind that, and one of the challenges that working with AWS has helped the museum to tackle, is around data management. Typically with research and education programs, the museum has used bespoke data solutions for each project, explains Tweddle: 

That comes with inefficiencies. It's hard to work collaboratively on projects, and the value of data scales up if you can bring it together. This is giving us a consistent platform we can use for all of our UK-based natural history, science and education going forward. That's a real long-term win.

As the museum combines different data streams to build a holistic view of the environment and wildlife, and as users enter more data in the system, Tweddle is expecting some bugs in the code to appear: 

We'll be testing those assumptions as the data goes in to see if we have got the right infrastructure and data flow, to enable us to accurately and quite carefully pull together these different forms of data so we can analyze them.

Data-rich

From being data-poor, thanks to advances in technology, the museum is now flooded with information on different species, their numbers and locations, and how the populations are changing. According to Tweddle: 

We've suddenly got so much data that in theory can give us this deep picture of the environment, which we can feed into understanding how changes we make do and don't benefit wildlife. We can use it to track progress towards biodiversity targets, but you have to pull that data together.

This is where cutting-edge scientists are looking for ways to combine these data sources, package up the data, and communicate it to the people who matter, like policy makers: 

This is our first test of that, which is why I'm saying there might be some bits of the code we need to tweak. It's very much an R&D kind of science.

Tweddle would like to see the platform opened up to more people in future, to use as the research hub for their own nature-based projects, as well as providing rapid access to science tools the museum is developing and to biodiversity metrics to understand how the changes they make are impacting nature:

That's one of our big science-led engagement and positive action goals. This infrastructure means that we can reasonably think about doing what we couldn't before.

While this project is UK-focused, the data system could eventually have wider impacts. Some of the smart tools embedded within the software are globally relevant, for example, naming systems for species that vary depending on the country. Tweddle argues: 

We've got a piece of machine-learning, graph database software, which allows us to intelligently match across the different naming systems, so we know we're comparing like for like when it comes to biodiversity data and how we share it. There are tools like that that can be globally relevant. A lot of the biodiversity data work we do is looking at global trends in wildlife and supportive businesses around sustainability at a global level. All the data we gather will be shared globally.

The next milestone for the museum will be the new garden galleries and biologically diverse green space, which will open in spring 2024. The five-acre site in South Kensington will have a broad network of sensors monitoring everything from the soil to UV levels and photosynthesis rates in trees. Tweddle says: 

The museum will be monitoring wildlife there and creating a digital representation in different ways. We can put a lot of tech into that area to build this 3D view of how nature in the space is behaving.

While the museum is focused on bringing new initiatives to life, like the Urban Nature Project, it’s also working on new ways to extract information from older collections. This includes exploring the use of AI and robotics to speed up the digitization process of its 80 million-plus specimens and provide increasingly nuanced data. Of the roughly five and a half million items that have been digitized so far, that data has been downloaded over 300 billion times across 2,500 science papers. Tweddle concludes: 

A whole data set from our digitized collection is downloaded every three minutes. There’s a huge benefit of extracting the data from our collections and that's where the tech comes in, to scale and speed that up because it's very labor intensive. But we know there's a huge value for the economy, to business, to nature of doing that.