Something bad is happening. Is it an accident or is there a more serious problem?
Hypothesis test and probability
Determine results from samples, surveys, and tests
Given existing experience, can we classify future events?
Clustering
Centroids of set of fields
Clustering is a form of classification
Group the data by minimizing some dissimilarity measure and assume data within group have same classification
K-means is popular unsupervised clustering algorithm
Requires two inputs: desired number of clusters $k$ and initial cluster centroids
Example of k-means clusteringCentroids colored by cluster
Can anyone catch the "error" in using this method?
Classification
Sample precision ag heatmap showing soybean yieldSimulated heatmap of field with irrigation pivotSimulated heatmap with no sign of irrigation
How might we go about solving this?
What makes the irrigated different from the non-irrigated?
How is a field typically planted?
If the path goes back-and-forth, what might we expect?
Could we not view the path as a time series?
Example of irrigated time series dataExample of non-irrigated time series dataHistogram of test statistics
Conclusion
There are many jobs out there other than teaching
People will pay you to solve problems all day
A degree in math is only a toolset, you will have to get creative with it