I'm not sure if this is related to #7413

Bug report

Example:

%pylab
x = randn(1000) / 1e5
y = randn(1000) / 1e5
scatter(x, y)

produces this:

It should really produce something more like this:

There's not an obvious way to do this, other than manually getting the min/max x and y values and adding a little bit and using ax.set_x/ylim()

Using conda on kubuntu:

matplotlib                2.1.0            py36hba5de38_0  
python                    3.6.3                h0ef2715_3