In our eating example, we may record how much we eat for a whole week and then make a note of how full we feel afterwards. But what's the point? The reason is to apply this knowledge in a meaningful way to help predict what will happen next. R code Knowing about how two things change together is the first step to predictionīeing able to describe what is going on in our previous examples is great and all. An exaggerated plot of no correlation between weight gain and test scores. For example, if you were to gain weight and looked at how your test scores changed, there probably won't be any general pattern of change in your test scores. There is also a third possible way two things can 'change'.

The faster the car, less travel time (trend to the bottom right). Negative correlation between car speed and travel time. This is a case of two things changing in the opposite direction (more speed, but less time). When you're in a car and it goes faster, you will probably get to your destination faster and your total travel time will be less. More food is eaten, the more full you might feel (trend to the top right).

Positive correlation between food eaten and feeling full. One goes up (eating more food), then the other also goes up (feeling full). This is a case of when two things are changing together in the same way. Here are some examples of the three general categories of correlation.Īs you eat more food, you will probably end up feeling more full.

A correlation is about how two things change with each otherĬorrelation is an abstract math concept, but you probably already have an idea about what it means. After reading this, you should understand what correlation is, how to think about correlations in your own work, and code up a minimal implementation to calculate correlations.

Correlations are a great tool for learning about how one thing changes with another.

R&r studio