Agrimetrics’ new Data Marketplace

Agrimetrics has recently launched a new Data Marketplace, which promises to reward data-owners whilst accelerating innovation.  Dr Matthew Smith, Agrimetrics’ Chief Product Officer, answers our questions about what it can offer farmers and others. 

Q. What type of data are you looking for and how should it be supplied – eg. could it come from individual farmers?

As our customers are both data providers and consumers, we’re interested in any data through which value could be created through its exchange.

Data could come from individual farmers in principle but it’s important to have an idea of a supply-demand relationship for the data, as we want to have confidence that there will be demand for data put in the marketplace.

If just one farmer puts up their data then it’d be unlikely to be attractive enough, but if it is part of a  larger data offering, e.g. data from a cooperative, (while importantly ensuring that the farmer is always in control of access to their data), then that could be interesting.

Another way to add value could be to link a farmer’s data to produce in the value-chain in order to support the reporting and certification of product provenance, which consumers are increasingly looking for.

That said, we also have strategic interests driven by where we think the priority needs are. In particular, data that enables:

• a view of the whole-farm
• agri-landscape and farm business sustainability
• cross supply chain data
• data relating to agri-food supply chain sustainability.

We’ll be announcing our public product roadmap shortly that’ll indicate some of our data acquisition priorities.

If we see value in the data then we will accept it in any form – in the traditional ways as files for us to upload or as models that generate the data – or the modern way as a direct transfer through the cloud.  A good example of this is the Airbus Verde dataset which generates 15 field attributes from satellite data and is available through our marketplace. When a customer requests some data we automatically pass the request on to Airbus who send the data back through us.

So why not just go to Airbus?

You can! But we add extra benefits: for example many people don’t just want the Verde data but also want other data related to that such as weather and soil data from other datasets in our marketplace that can be combined with the Verde data to create new insight.

The point is, right now we’re putting a lot of thought into finding viable data exchanges for which there is a willing provider and consumer.

Q. How will data providers be paid for their data? Is there an annual fee or per usage? Will the data be anonymised?

Whatever the data provider wants! It’s access to their data that they’re selling.

However, as this is a marketplace, they’ll want to be sure the price and pricing mechanism that is right for the data consumer. On our product roadmap is providing a way to enable data providers to automatically set their price and payment mechanism.

We have a few different ways to enable data consumers to pay for data right now –  a subscription model and a pay-per-use model – depending on the datasets. However we’re open to considering other models.

The data does not need to be anonymised, again it’s up to the data provider if they want to, or it makes sense to, anonymise data,  we treat these things on a case by case basis.

Q. How can users access the data? Can you make a query, such as ‘tell me where the onion producers are in the UK?’ or do you need to be a data analysis professional?

We have a few different ways. Our main mode to enable data access is directly from our cloud databases over what’s called an API.

We have several APIs that support different kinds of query, for example one dedicated to facts about agricultural fields in the UK. We just launched a new one called the GraphQL API, which allows more complex queries about linked data to be made – so you could ask for ‘all the fields for which the sown-crops are onions’. We don’t currently have the details of the producers, just the fields.

You would need to be someone with programming skills to directly write the API queries however we have produced a number of other tools to make things easier.

Our Query Builder is a web-based graphical tool that allows you to select the data you want by navigating our knowledge graph and selecting what you want, and then the query gets built in the browser. From there you can either run the query in the browser or export the code that’s automatically written and embed it in a tool that can run the query like Excel.

Making this more accessible to inexperienced or non-programmers is a major area of work over the coming months. Critically, we are indeed aiming to grow the range of different types of query we can support.

Q. Will you include public datasets, i.e. information provided by farmers for regulatory purposes?

We include public datasets already, those that are made available for free from the likes of the Met Office. We’re definitely looking at how we support the exchange of data between farmers and regulators.

A key way we can help here is to give farmers a mechanism where they can grant access to particular types of data to users on a case-by-case basis. This will enable farmers to adopt a “measure once, use many” approach rather than the current unsustainable approaches of collecting or inputting effectively the same data multiple times to support multiple purposes.

Q. What are you hoping that it will achieve?

We set up the marketplace as a key way to realise our mission at Agrimetrics of accelerating the realisation of new value in the agri-food sector from data. For us it was important to identify a scalable way to serve the whole sector. We ultimately reduce the costs for both data providers and consumers in making data exchanges, enabling the value to flow faster, thus accelerating the industry.

All this is part of the solutions the UK and the world needs to our major sustainability and food security challenges – a lack of adoption of data-informed solutions is holding back the sector, meaning sustainability, productivity and growth improvements are not being made as fast as they need to be. Like our food, we need sustainable data supply chains!

Q. Can you give some tangible examples of the datasets that have already been contributed?

The biggest new dataset is the Verde Field Attributes dataset – others are datasets we’ve associated with UK fields such as Soil data from CEH and Weather data from the Met Office .

Take a look at our catalogue: app.agrimetrics.co.uk/#/catalog/data-sets