Improving clustering plots.

clustering/kde.py

@@ -1,4 +1,4 @@
-# Copyright 2015 Rodrigo Roim Ferreira
+# Copyright 2015 Rodrigo Roim Ferreira
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by

plotting/clustering.py

@@ -1,4 +1,4 @@
-# Copyright 2015 Rodrigo Roim Ferreira
+# Copyright 2015 Rodrigo Roim Ferreira
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -18,19 +18,34 @@
 import matplotlib.pyplot as _pl
 import numpy as _np
 
-import clustering as _clst
+import clustering.kde as _kde
 
 
-# Sample data (Ode to Joy): x = _np.array([5.129283017, 4, 5.044394119, 4, 4, 3.906890596, 4.857980995, 3.906890596, 5.129283017, 5.129283017, 5.087462841, 3.807354922, 5.044394119, 4, 4.087462841, 4, 4.700439718, 3.906890596, 4.087462841, 4, 4.857980995, 4.584962501, 3.906890596, 3.321928095, 3.807354922, 3.807354922, 3.807354922, 4.169925001, 2.807354922, 3.807354922, 3.906890596, 3.906890596, 4.169925001, 3.906890596, 3.700439718, 3.700439718, 5.129283017])
-def plot_kde(x, bw=0.15, bounds=(1,12)):
+# Sample data (Ode to Joy): x = _np.log2([20, 23, 21, 23, 23, 22, 24, 22, 22, 20, 22, 22, 34, 10, 44, 21, 21, 21, 23, 23, 22, 25, 22, 50, 22, 22, 33, 10, 50, 21, 21, 20, 25, 22, 10, 11, 25, 25, 22, 10, 10, 24, 22, 25, 22, 26, 44, 23, 20, 22, 22, 21, 20, 11, 10, 50, 22, 24, 33, 11, 53])
+def plot_histogram(x, bounds=(1, 12)):
+    """Plots a histogram for a given array."""
+    plot_width = bounds[1] - bounds[0]
+    hist_points = _np.linspace(bounds[0], bounds[1], 10*plot_width)
+    _pl.hist(x, bins=hist_points, range=bounds, fc='gray', histtype='stepfilled', alpha=0.5, normed=False)
+
+    _pl.xlabel("Duração (lg da quantidade de janelas)")
+    _pl.ylabel("Quantidade de notas")
+    _pl.show()
+    return
+
+
+def plot_kde(x, bw=0.15, bounds=(1, 12)):
     """Plots a histogram and the KDE for a given array."""
     plot_width = bounds[1] - bounds[0]
 
     hist_points = _np.linspace(bounds[0], bounds[1], 10*plot_width)
     _pl.hist(x, bins=hist_points, range=bounds, fc='gray', histtype='stepfilled', alpha=0.5, normed=True)
 
     kde_points = _np.linspace(bounds[0], bounds[1], 100*plot_width)
-    pdf = _clst.kde(x, kde_points, bw)
+    pdf = _kde.kde(x, kde_points, bw)
     _pl.plot(kde_points, pdf, color='red', alpha=0.75, lw=2)
 
+    _pl.xlabel("Duração (lg da quantidade de janelas)")
+    _pl.ylabel("Quantidade de notas (normalizada)")
     _pl.show()
+    return