added cumulative histograms

mmetz-bn · mmetz-bn · commit a7820920dd22 · 2008-05-16T14:52:15.000Z
svn path=/trunk/matplotlib/; revision=5147
diff --git a/CHANGELOG b/CHANGELOG
@@ -1,3 +1,7 @@
+2008-05-16 Added 'cumulative' keyword arg to hist to plot cumulative
+           histograms. For normed hists, this is normalized to one
+           assuming equally spaced bins - MM
+
 2008-05-15 Fix Tk backend segfault on some machines - MGD
 
 2008-05-14 Don't use stat on Windows (fixes font embedding problem) - MGD
diff --git a/examples/histogram_demo_cumulative.py b/examples/histogram_demo_cumulative.py
@@ -0,0 +1,33 @@
+#!/usr/bin/env python
+from pylab import *
+
+mu, sigma = 100, 25
+x = mu + sigma*randn(10000)
+
+# the histogram of the data
+n, bins, patches = hist(x, 50, normed=1, histtype='step', cumulative=True)
+setp(patches, 'facecolor', 'g', 'alpha', 0.75)
+
+# add a 'best fit' line
+y = normpdf( bins, mu, sigma).cumsum()
+y /= y[-1]
+l = plot(bins, y, 'k--', linewidth=1.5)
+
+# overlay the first histogram with a second one
+# were the data has a smaller standard deviation
+mu, sigma = 100, 10
+x = mu + sigma*randn(10000)
+
+n, bins, patches = hist(x, bins=bins, normed=1, histtype='step', cumulative=True)
+setp(patches, 'facecolor', 'r', 'alpha', 0.25)
+
+# add a 'best fit' line
+y = normpdf( bins, mu, sigma).cumsum()
+y /= y[-1]
+l = plot(bins, y, 'k--', linewidth=1.5)
+
+grid(True)
+ylim(0, 1.1)
+
+#savefig('histogram_demo',dpi=72)
+show()
diff --git a/lib/matplotlib/axes.py b/lib/matplotlib/axes.py
@@ -5416,9 +5416,9 @@ def twiny(self):
     #### Data analysis
 
 
-    def hist(self, x, bins=10, normed=False, bottom=None, histtype='bar',
-             align='edge', orientation='vertical', width=None,
-             log=False, **kwargs):
+    def hist(self, x, bins=10, normed=False, cumulative=False,
+             bottom=None, histtype='bar', align='edge',
+             orientation='vertical', width=None, log=False, **kwargs):
         """
         HIST(x, bins=10, normed=False, bottom=None, histtype='bar',
              align='edge', orientation='vertical', width=None,
@@ -5439,6 +5439,12 @@ def hist(self, x, bins=10, normed=False, bottom=None, histtype='bar',
           pdf, bins, patches = ax.hist(...)
           print np.trapz(pdf, bins)
 
+        If cumulative is True then histogram is computed where each bin
+        gives the counts in that bin plus all bins for smaller values.
+        The last bins gives the total number of datapoints.  If normed is
+        also True then the histogram is normalized such that the last bin
+        equals one (assuming equally spaced bins).
+
         histtype = 'bar' | 'step'. The type of histogram to draw.
         'bar' is a traditional bar-type histogram, 'step' generates
         a lineplot.
@@ -5461,6 +5467,12 @@ def hist(self, x, bins=10, normed=False, bottom=None, histtype='bar',
         if not self._hold: self.cla()
         n, bins = np.histogram(x, bins, range=None,
             normed=bool(normed), new=True)
+        
+        if cumulative:
+            n = n.cumsum()
+            if normed:
+                # normalize to 1
+                n *= (bins[1]-bins[0])
 
         if histtype == 'bar':
             if width is None:
