In association with the HGCA's BePRECISE initiative

In the final academy on precision farming, BePRECISE's Ian Beecher-Jones advises on how to get value from the data you've generated.

Precision farming is incredibly data rich. Used correctly, information can benefit a business; left on a memory stick or in a cupboard it is a waste of money.

By knowing what you want to do with the data before you start, you can save a huge amount of time and confusion after they have been collected.

If you can measure it, you can manage it

By measuring different aspects of your farming operation you will be able to make business decisions about how you manage individual fields. For example, you may know you have high and low yielding areas in your fields.

By testing for P & K you can work out whether P & K could be resulting in lower yields and then provide remedial treatment to change the indices. Without measuring the problem, you will not be able to manage it.

Other than trying to quantify the variability across a field, there are other uses for data you collect that will improve machine and operator performance.

For example, with the support of telemetry, you can recognise good practice and share it with other members of staff. However, as there is the ability to collect so much data, be sure what you want to measure and why. Otherwise you may become bogged down by so much detail that it becomes unmanageable.

It’s the human touch

Some believe the most important part of data management is the computer or the machine – it isn’t. It is you the farmer.

Without your time and dedication it will become a costly burden. With your understanding and engagement the farm will profit from your informed and accurate decisions.

Before you start you have to have clear objectives on data from:

What type of data can you collect?

Data can be collected from many different freely available sources (see data-gathering techniques, table 1) and can be very useful in understanding variation in a ?field and recognising the way ?fields react.

The difference between traditional methods and precision farming methods is that each point is assigned a georeferenced point to know exactly where a particular value is located.

What do data look like?

Data are perceptively very complex. However, in the raw form they are very simple. All have common attributes – a GPS position (latitude and longitude) and a value or multiple values:

51.47361   -1.39158  10.08

All the other numbers in the comma-separated file format (CSV) are values, which could include moisture, number of satellites, speed, etc. The raw information is the same, but can be stored in many different formats depending on the device, manufacturer or software provider.

What are good data?

Good data are useful, bad data are not. Good data should be:

• Timely.
• Relevant to the application.
• Captured from a well-calibrated and correctly set machine.
• Where human or mechanical errors are minimised.
• Data which have the potential to provide a return on investment.
• Able to identify treatable problems.

Collecting useful data doesn‘t happen by accident, care has to be given to their collection. One of the most important aspects is the calibration of equipment. If the equipment isn’t calibrated, then you may be spreading precisely, but inaccurately. An incorrectly calibrated machine won’t magically provide reliable data.

Getting prepared

Before starting any operation ensure that the operator and machine understand each other. For example, ensure a combine is set up correctly to match the driver’s particular style rather than the other way round; this will ensure consistent working across the field and remove any potential errors such as altering the cutter bar width on narrow runs. If not, errors can be created in the maps which are costly to remove.

It’s all about scale

The perception is that highly detailed maps are the ideal of precision farming. Generally this is not the case. Trends across a field are much more useful, as they will relate directly to your piece of application equipment.

If you have a 36m wide sprayer with auto section control every 3m, why try to manage the data down to 1m – the sprayer can’t discriminate that level of detail.

A number of software providers allow maps to be linked to services such as Google Earth so you can visualise the trends more clearly.

How are data managed?

There are a number of different routes to handling data. The diagram (right) highlights the interaction between, at one end, a farmer handling the whole process and at the other a service provider managing the process with little farmer ?involvement.

Each method has its advantages and disadvantages (table 2) and it will be up to each farmer to decide how to handle it. Either way the farmer has to understand what is happening and why you are doing what you are doing. Randomly asking data to be collected will serve no useful purpose.

What do I do with all the data when I have got them?

You should have an answer to that question before the data are collected; recreational collection of data is expensive and inadvisable.

One end result of collecting good data and having a plan for dealing with them would be the creation of an application plan and map for application of seeds, fertilisers and sprays working in conjunction with your agronomist and guides such as RB209. While you are creating the application plan you will have to be aware of the range of different file formats which are required by different manufacturers. Being aware of these differences and the ability of your software to handle them is essential to ensure the entire system will work.