LIC – Disrupting dairying with dataviz

Herd test results

Normally dairy farmers provide their milk in bulk, whereupon the dairy company checks it for “good” and “bad” contents. A herd test is when samples are taken from individual cows, to see how they are affecting those average amounts.

Prior to 2011, LIC‘s farmer customers received their herd test results as a paper printout of tabular figures spanning many pages. When it came time to move that information to the MINDA Web system, I suggested a rethink.

Following the work of Stephen Few and others, I tried to apply modern data visualisation principles to the problem. Talking to real farmers, I quickly learned that extracting meaningful information out of the printed report required much use of pencils, highlighters and rulers to shortlist both “good” and “bad” cows, as well as frequent reference to earlier reports as there was no trend information.

I asked farmers what actions farmers would take based on their herd test reports, and came up with three:

  • Breed from this cow – she has good genetics and she is producing good milk
  • Treat this cow – she has good genetics but her milk is poor quality, suggesting a health issue
  • Remove this cow – she has poor genetics and she is producing poor milk, making her a liability

In this graph, the x axis represents milk quality and the Y axis represents the genetic potential of the cow. Every dot represents a cow. If a cow falls in one of the two left hand quadrants, she likely requires intervention from the farmer. The farmer can ignore all the cows in the right hand quadrants as they are doing fine.

Heifer rearing

When farmers select the newborn female calves that will be their future milking herd, the calves are usually sent to a rearer for the next two years. If a two year old heifer is underweight at the end of this process, it will impact her reproductive performance for the rest of her life.

Traditionally, farmers have little to no visibility of how their calves are performing when they are away at the rearers.

In 2012, we added MINDA Weights, a web product that is designed to encourage calf rearers to use electronic scales to provide regular updates to their farmer clients.

Again, I began by asking calf rearers what actions they would take based on weighing the young stock.

  • Action – If a calf is both underweight and not gaining fast enough, she goes into a special group to receive extra feed and supplements
  • Monitor – If a calf is at her target weight but she is gaining slower than she should be, place her into a group that gets regular checks
  • Recovering – If a calf is underweight but she is gaining at a good rate, she can be moved back out of the special group

In this graph, the x axis represents the rate of weight gain and the y axis represents the rate of gain. The quadrant that a calf is in represents the management group (or paddock) she should be in.

In addition to guidance on actions, calf rearers needed a quick way to tell if the overall group of calves they just weighed was ok or in need of attention.

In this graph, a red highlight and warning message appears if more than 10% of the calves are underweight. The warning is only applied to the most recent weighing, as alerting a user about historical data is “crying wolf”.