How to enter manually a key-value to seaborn jointplot stat_func?

Question

i want to display some parameter on a seaborn jointplot.

Lets say apples=5 just like pearson=.3

I am not interested in the default option. so i generated the graph with the following line of code:

sns.jointplot(sp.time, mn, color="#4CB391", stat_func=None)

The Documentation states:

stat_func : callable or None, optional
Function used to calculate a statistic about the relationship and annotate the plot. 
Should map x and y either to a single value or to a (value, p) tuple. 
Set to None if you don’t want to annotate the plot.

could someone help fill in the stat_func properly do display a key-value pair of my choosing?

Thank you.

The plot is

Answer 1

You can create you own function that takes two input parameters (x and y from the call to sns.jointplot()) and returns a value or a tuple of two values. If you just want to display arbitrary text, it is better to use ax.text() as indicated in @mwaskom's comment to your question. But if you are calculating something from a function of your own, you could do:

import seaborn as sns
tips = sns.load_dataset('tips')

def apples(x, y):
    # Actual calculations go here
    return 5

sns.jointplot('tip', 'total_bill', data=tips, stat_func=apples)

If apples() returned two values in a tuple (e.g. return (5, 0.3)), the second value would represent p and the resulting text annotation would be apples = 5; p = 0.3.

You could compute any statistics like this, as long as the return format is a single value or a (value, p) tuple. If you want to use scipy.stats.kendalltau for example, you would do

from scipy import stats
sns.jointplot('tip', 'total_bill', data=tips, stat_func=stats.kendalltau)