turn scatter data into binned data with errors bars equal to standard deviation

Question

I have a bunch of data scattered x, y. If I want to bin these according to x and put error bars equal to the standard deviation on them, how would I go about doing that?

The only I know of in python is to loop over the data in x and group them according to bins (max(X)-min(X)/nbins) then loop over those blocks to find the std. I'm sure there are faster ways of doing this with numpy.

I want it to look similar to "vert symmetric" in: http://matplotlib.org/examples/pylab_examples/errorbar_demo.html

Solution

You can bin your data with np.histogram. I'm reusing code from this other answer to calculate the mean and standard deviation of the binned y:

import numpy as np
import matplotlib.pyplot as plt

x = np.random.rand(100)
y = np.sin(2*np.pi*x) + 2 * x * (np.random.rand(100)-0.5)
nbins = 10

n, _ = np.histogram(x, bins=nbins)
sy, _ = np.histogram(x, bins=nbins, weights=y)
sy2, _ = np.histogram(x, bins=nbins, weights=y*y)
mean = sy / n
std = np.sqrt(sy2/n - mean*mean)

plt.plot(x, y, 'bo')
plt.errorbar((_[1:] + _[:-1])/2, mean, yerr=std, fmt='r-')
plt.show()