15.1

Schrick-Noah_Learning-Practice-4.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 55,
+   "execution_count": 122,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -12,9 +12,13 @@
     "# Noah L. Schrick - 1492657\n",
     "\n",
     "import pandas as pd\n",
+    "import numpy as np\n",
     "from sklearn.metrics import pairwise\n",
     "import matplotlib.pylab as plt\n",
     "from scipy.cluster.hierarchy import dendrogram, linkage, fcluster\n",
+    "from tabulate import tabulate\n",
+    "from scipy import cluster\n",
+    "import seaborn as sns\n",
     "\n",
     "pd.options.mode.chained_assignment = None  # default='warn'\n"
    ]
@@ -46,7 +50,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 60,
+   "execution_count": 134,
    "metadata": {},
    "outputs": [
     {
@@ -58,6 +62,15 @@
      },
      "metadata": {},
      "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Colorado Christian University\n",
+      "475   NaN\n",
+      "Name: # PT undergrad, dtype: float64\n"
+     ]
     }
    ],
    "source": [
@@ -77,6 +90,8 @@
     "                                     metric='euclidean')\n",
     "pd.DataFrame(university_dist, columns=university_df.index, index=university_df.index).head(5)\n",
     "\n",
+    "uni_hclust = fcluster(linkage(university_df_num_norm, 'complete'), 6, criterion='maxclust')\n",
+    "\n",
     "Z = linkage(university_df_num_norm, method='complete')\n",
     "\n",
     "fig = plt.figure(figsize=(10, 6))\n",
@@ -88,7 +103,36 @@
     "plt.xticks(rotation=45, ha='right')\n",
     "plt.show()\n",
     "# reasonable number of clusters for describing the data:\n",
-    "# At distance of 10.5 (horizontal line in the dendrogram image) data can be reduced to 9 clusters"
+    "# At distance of 10.5 (horizontal line in the dendrogram image) data can be reduced to 9 clusters\n",
+    "\n",
+    "# c:  Compare the summary statistics for each cluster\n",
+    "cutree = cluster.hierarchy.cut_tree(Z, n_clusters=[5, 10])\n",
+    "clust_stats = university_df_num_norm.agg(['mean', 'median'])\n",
+    "\n",
+    "# d: Use the categorical measurements to categorize\n",
+    "state_table = tabulate(university_df[['State']], cutree)\n",
+    "pub_priv_table = tabulate(university_df[['Public (1)/ Private (2)']], cutree)\n",
+    "\n",
+    "#university_df_num_norm.index = ['{}: {}'.format(cluster, state) for cluster, state in zip(uni_hclust, university_df.index)]\n",
+    "#sns.clustermap(university_df_num_norm, method='average', col_cluster=False,  cmap=\"mako_r\")\n",
+    "#plt.show()\n",
+    "\n",
+    "# e: Other external information\n",
+    "# Text answer\n",
+    "\n",
+    "# f: Compute the Euclidean distance of this record from each of the clusters that you found above (using only the measurements that you have)\n",
+    "tufts_df = raw_university_df.loc[raw_university_df['College Name'] == 'Tufts University']\n",
+    "tufts_df = tufts_df.drop(['# PT undergrad'], axis=1)\n",
+    "tufts_df_num = tufts_df.select_dtypes(include='number') # get numeric cols only\n",
+    "\n",
+    "tufts_dist = pairwise.pairwise_distances(tufts_df_num, Y=university_df_num_norm, metric='euclidean')\n",
+    "\n",
+    "# Closest cluster:\n",
+    "print(raw_university_df.iloc[np.where(tufts_dist == tufts_dist.min())[1][0]]['College Name'])\n",
+    "\n",
+    "# impute missing (from raw data - non-normalized)\n",
+    "tufts_df['# PT undergrad'] = clust_stats['# PT undergrad']['mean']\n",
+    "print(tufts_df['# PT undergrad'])"
    ]
   },
   {
@@ -125,10 +169,23 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 136,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
-   "source": []
+    {
+     "data": {
+      "text/plain": [
+       "3.771458257750001e-17"
+      ]
+     },
+     "execution_count": 136,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "clust_stats['# PT undergrad']['mean']"
+   ]
   }
  ],
  "metadata": {