Path to Precision

Action Plan Step 3: Evaluate Data Layers

13 Dec 2021

Once you’ve got someone taking care of your precision farming strategy, and you’ve assessed what precision farming data and hardware you need and have access to, the next step is about the software and the data captured from that hardware—in other words, evaluating your data layers.

First let's clarify what is meant by a 'data layer'. When we use data in precision farming, we use them in layers—think layers of a cake.  

Each layer shows the measurements of a particular parameter across a recorded location. In agronomic terms this could be a yield layer, a soil layer, an elevation layer, or an NDVI (Normalised Difference Vegetation Index) crop health layer. And for a machine it could be a fuel layer, speed layer, engine load layer and many more! The FMIS will then use the layers, stack them on top of each other and draw its conclusions.

So you need to know what you have, the quality of what you have, and what you need.

As we've discussed earlier in Path to Precision, to do anything with the data captured, you need a Farm Management Information System (FMIS). And that’s the case with this step. However, this doesn't mean paying a fee and locking yourself into an FMIS before you know what you're doing! This is a good opportunity to start learning about Farm Management Information Systems and get familiar with the options.

Most manufacturers of precision ag equipment will have a free version of their FMIS platform that allows you to visualise data, i.e., visualise your layers. Case IH is no different. AFS Connect is free to use for anyone with some data—you can sign yourself up today! Use these free versions of software to assess your data layers.

You should have some idea of the layers available to you from the previous step of assessing inventory. ​Examples of extracting the data layers may be from a USB in a vehicle display, from an agronomist, a satellite remote sensing platform or even a high-tech drone you may have available. Whatever you have, load the data into your free version of software and see how it all looks. Make sure you are measuring what you think you are measuring!
You should be checking things like:

Data quality - colours are usually a good indicator. If you have a large area of red in the middle of your yield data layer (i.e. your field) then this could indicate a poor calibration or defective sensor.
Data gaps - or missing data. You should have a full layer of data across your fields. Holes in your data could indicate poor sensor, poor calibration or poor operation.
General data 'eyeballing' - use your common sense and expert knowledge about your crops and farm and match it with the data. Does it seem right? Does it make sense? Did you really spread 100 tonnes/ha of potash in that area? Investigate anything that doesn't add up—it could be a simple data entry mistake or again point to a sensor issue.

Learn to import and read data, how to switch between data layers, get comfortable with the process and importantly, learn what your farm performance benchmark is.

The step of evaluating your data layers is complete when:

1. You have the data layers you need for your precision strategy goals
2. Those data layers are 'healthy' and the quality of them is good
3. You have familiarised yourself with the data handling process of your operation
Next, the big one... Step 4: Evaluate Yield!