Defending Dignity A Manual for National Human Rights Institutions on Monitoring Economic, Social and Cultural Rights The relationship between the two variables is called their correlation. When the two data sets are strongly linked, it can be said that they are “highly correlated”. Correlation is positive when the values increase together, and correlation is negative when one value decreases as the other increases. In this example there is a negative correlation between the two variables; more spending on teachers’ wages is associated with lower student-teacher ratios. You can also add a trend line to the scatter plot to make the negative/positive relationship clear for the reader. This is particularly useful when you have only few data points. The R2 coefficient provides a statistical estimate of how closely the data points are correlated. An R2 of 1 indicates that the two data points are perfectly correlated; an R2 close to 0 indicates that there is very little relationship between the two data points. 15.3. BASIC DESIGN PRINCIPLES The effectiveness of data visualization can be judged according to its: • Appeal: It should attract the reader’s attention • Comprehension: It should enable clear understanding of the information, so the reader can get meaning from it • Retention: It should be memorable, so that the reader takes away a key message. The weight given to each of these factors will depend on the objective you are trying to achieve (discussed further in Chapter 17). For example, if you are presenting your findings to a parliamentary committee, it is important that the information is seen as clear and unbiased. As such, comprehension would be the primary goal. On the other hand, if the data is being used to support a public awareness campaign, retention might be the key priority because you want people to share what they have learned with others. Do not to manipulate data into a pre-defined narrative. Bad visualization can harm an advocacy message, draw attention away from the main point or, in the worst case, become misleading. Below are some basic principles to follow for accurate and effective data visualization. 15.3.1. Size matters In data visualization, the size of the axis really does matter. For example, the two column charts here show the percentage of vulnerable households that consumed selected food items weekly in 2012, compared to 2013. PERCENTAGE OF VULNERABLE HOUSEHOLDS THAT CONSUME SELECTED FOOD ITEMS WEEKLY 2012–2013 100 100 95 90 85 95 80 75 70 90 65 60 55 85 Legumes Fruit 2nd quarter 2012 Source: Hypothetical data 150 Proteins Dairy products 2nd quarter 2013 50 Legumes Fruit 2nd quarter 2012 Proteins Dairy products 2nd quarter 2013