Use np.histogram to generate the counts, divide these by the sum of

the counts, and use pyplot.bar to plot the bar heights. Something

like

In [44]: x = np.random.randn(10000)

In [45]: n, bins = np.histogram(x, bins=20)

In [46]: left = bins[:-1]

In [47]: width = bins[1] - bins[0]

In [48]: pct = n/float(n.sum())

In [49]: plt.bar(left, pct, width=width)

JDH

## ยทยทยท

On Mon, Dec 14, 2009 at 10:22 AM, Susanne Pfeifer <tiffy@...2899...> wrote:

Hello,

I am relatively new to matplotlib and I was wondering whether there is

an easy possibility to generate a histogram whose height is normalized

to one (rather than the total area under the curve which is the case if

I use normed=1).