ready for binder

classification_utilities.py

@@ -12,42 +12,50 @@ def display_cm(cm, labels, hide_zeros=False,
     recall = np.diagonal(cm)/cm.sum(axis=1).astype('float')
     F1 = 2 * (precision * recall) / (precision + recall)
 
+    precision[np.isnan(precision)] = 0
+    recall[np.isnan(recall)] = 0
+    F1[np.isnan(F1)] = 0
+
+    total_precision = np.sum(precision * cm.sum(axis=1)) / cm.sum(axis=(0,1))
+    total_recall = np.sum(recall * cm.sum(axis=1)) / cm.sum(axis=(0,1))
+    total_F1 = np.sum(F1 * cm.sum(axis=1)) / cm.sum(axis=(0,1))
+    #print total_precision
+
     columnwidth = max([len(x) for x in labels]+[5]) # 5 is value length
     empty_cell = " " * columnwidth
     # Print header
-    print "    " + " Pred",
+    print("    " + " Pred", end=' ')
     for label in labels: 
-        print "%{0}s".format(columnwidth) % label,
-    print
-    print "    " + " True"
+        print("%{0}s".format(columnwidth) % label, end=' ')
+    print("%{0}s".format(columnwidth) % 'Total')
+    print("    " + " True")
     # Print rows
     for i, label1 in enumerate(labels):
-        print "    %{0}s".format(columnwidth) % label1,
+        print("    %{0}s".format(columnwidth) % label1, end=' ')
         for j in range(len(labels)): 
             cell = "%{0}d".format(columnwidth) % cm[i, j]
             if hide_zeros:
                 cell = cell if float(cm[i, j]) != 0 else empty_cell
-            print cell,
-
-        print
+            print(cell, end=' ')
+        print("%{0}d".format(columnwidth) % sum(cm[i,:]))
 
     if display_metrics:
-        print
-        print "Precision",
+        print()
+        print("Precision", end=' ')
         for j in range(len(labels)):
             cell = "%{0}.2f".format(columnwidth) % precision[j]
-            print cell,
-        print
-        print "   Recall",
+            print(cell, end=' ')
+        print("%{0}.2f".format(columnwidth) % total_precision)
+        print("   Recall", end=' ')
         for j in range(len(labels)):
             cell = "%{0}.2f".format(columnwidth) % recall[j]
-            print cell,
-        print
-        print "       F1",
+            print(cell, end=' ')
+        print("%{0}.2f".format(columnwidth) % total_recall)
+        print("       F1", end=' ')
         for j in range(len(labels)):
             cell = "%{0}.2f".format(columnwidth) % F1[j]
-            print cell,
-        print
+            print(cell, end=' ')
+        print("%{0}.2f".format(columnwidth) % total_F1)
 
 
 def display_adj_cm(

environment.yml

@@ -2,6 +2,7 @@ dependencies:
 - python=3.5.2
 - numpy
 - matplotlib
+- seaborn
 - jupyter
 - pandas
 - scikit-learn

index.ipynb

@@ -28,7 +28,9 @@
     "\n",
     "All the code and data to reproduce *everything* in that article is right here in this repository. You can read the code in a [Jupyter Notebook](http://jupyter.org/) here...\n",
     "\n",
+    "<div style=\"width:50%; margin: 12px 0px 6px 20px; padding: 8px; border: 2px solid darkblue; border-radius: 6px; font-size: 125%; background: #EEEEFF;\">\n",
     "[**Facies_classification.ipynb**](Facies_classification.ipynb)\n",
+    "</div>\n",
     "\n",
     "See [the February issue of *The Leading Edge*](http://library.seg.org/doi/abs/10.1190/tle35020190.1) for Matt Hall's user guide to the tutorials; it explains how to run a Jupyter Notebook.\n",
     "\n",
@@ -77,15 +79,16 @@
   }
  ],
  "metadata": {
-      "kernelspec": {
-        "display_name": "Python [conda env:binder]",
-        "language": "python",
-        "name": "conda-env-binder-py"
- },
-   "language_info": {
-     "codemirror_mode": {
-      "name": "ipython",
-      "version": 3
+  "anaconda-cloud": {},
+"kernelspec": {
+      "display_name": "Python [conda env:binder]",
+      "language": "python",
+      "name": "conda-env-binder-py"
+    },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
    },
    "file_extension": ".py",
    "mimetype": "text/x-python",