Hello all!

Do you have an idea how best visualise the data from this table knowing they are coefficients of binomial logistic regressions? What I would like to visualise is a confrontation between the predictors likely to influence 'paying intent'

Also what is the difference is between 'beta' and 'exponential(beta)'?

edit: All of the variables are binomial but for the age variable that is continuous.

## Best Answer

With binary predictors and a binary outcome, there are only 4 cells (conditions, or possibilities) to display: predictor = either 0 or 1 and outcome = either 0 or 1. Thus one can use a very simple format like a table or bar chart to show the connections between the two variables.

With a continuous predictor, the relationship may be more complex. One effective way to display this is through a sunflower plot.

Each circle represents one case, and each sunflower "petal" represents an additional case that has the same Predictor and Outcome values as the rest of the cases in that circle.

With such a chart one can see exactly how many cases have each pair of values; to what extent the relationship departs from a linear one; and where the slope discriminating between "No" and "Yes" outcomes is steepest.

As for 'beta' and 'exponential(beta)', any general source on logistic regression will explain the difference, but in short: with the latter one raises beta to a power (the coefficient), base

e. This yields the odds ratio associated with a case being 1 higher than another on the predictor. For example, when Music is '1' the odds of the outcome being '1' are 1.6 times the odds when Music is '0'.