Upcoming CNG event

Virtual Conference 2024

13 November 2024

Creating Interoperable Field Boundary Data with the fiboa Converter Tool

In May, we discussed Field Boundaries for Agriculture (fiboa) and the fiboa ecosystem and mentioned that there is a new converter tool, which can take non-fiboa datasets and help you turn it into fiboa datasets. Back then we had 5 very similar datasets converted. In the meantime, we’ve converted additional datasets and improved the converter tool. Today, we’d like to give an update on the status and show how easy it is for you to make your field boundaries more useful by converting and providing them in a “standardized” format.

Seven people are currently working on creating more than 40 converters:

How does it work that we can convert so many datasets so easily?

We have implemented the fiboa Command Line Interface (CLI), which is a program that offers various tools to work with field boundary data. One of them is a command to convert field boundary datasets from their original form into fiboa. As field boundary data usually looks very similar conceptually (a geometry plus additional properties), most of the steps that are needed to convert the datasets to a standardized format can be abstracted away.

This means that the converter just needs a couple of instructions that describe the source data and then it can do all other steps automatically. The code is in Python, but we provide a template that is documented and contains examples. It’s pretty simple to fill and most people that have some programming experience should be able to make the necessary changes. Let’s look at the most important steps in the conversion process:

Additionally, you should provide some general metadata about the dataset (e.g., title, description, provider, license) so that people know what they can expect. That’s not really needed for the conversion though so I don’t cover them here. The template also provides additional options and parameters that can be fine-tunes for your needs. You can have a look at the full template. It has documentation included and gives some examples. For even more examples check out the filled templates for the existing converters. There’s also a tutorial that describes how to create a converter, either in written form or as a YouTube video.

Once this is done, you can make the converter available by creating a Pull Request on GitHub and eventually it will be available to others. Everyone with access to the data can then convert to fiboa at any time, e.g. if the source data has been updated. It’s really simple. To convert data you just can use the following command:

fiboa convert X -o result.parquet

Just replace X with any of the available converters. Use the command fiboa converters to list all available converters.

The result.parquet file can then be loaded into any software that can read Parquet files, for example QGIS. It can also be loaded in many programming languages such as Python or R. Loading multiple fiboa-compliant datasets makes it much simpler to work across multiple datasets as many properties are already aligned, so it’s analysis-ready and can be used very fast and efficiently.

This makes it very easy to work with massive amounts of field boundary data, even across providers, and work against a standardized interface without the need to manually preprocess the data heavily to be comparable. For more details see also the section “Why a farm field boundary data schema?” in our blog post “Introduction to fiboa”.

Join us and make your datasets available in a standardized format now! We are happy to help if you run into issues, just get in touch via email or GitHub issues or pull requests.


Our blog is open source. You can suggest edits on GitHub.


Connect with us: