{
 "metadata": {
  "name": "",
  "signature": "sha256:d5e7185920adf0ffeb2871081f81d27b79eaff52c3f3976f0c432d0a5020d6d0"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# NumPy Data Access Using ArcPy "
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import arcpy as ARCPY\n",
      "import arcpy.da as DA\n",
      "inputFC = r'../data/CA_Polygons.shp'\n",
      "tab = DA.TableToNumPyArray(inputFC, ['PCR2000', 'POP2000'])\n",
      "print(tab)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# SSDataObject\n",
      "1. Environment Settings (Except Extent)\n",
      "2. Bad Records\n",
      "3. Error/Warning Messages\n",
      "4. Localization\n",
      "5. **Feature Accounting**\n",
      "  * Cursors and DataAccess are not assured to read attributes in order.\n",
      "\n",
      "  * Keeps track of the shapes and their attributes so that one can create output features w/o post-joins.\n",
      "\n",
      "  * Unique ID works with Spatial Weights Formats in ArcGIS, PySAL, R, Matlab, GeoDa etc.."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import SSDataObject as SSDO\n",
      "ssdo = SSDO.SSDataObject(inputFC)\n",
      "ssdo.obtainData(ssdo.oidName, ['PCR2000', 'POP2000'])\n",
      "print(ssdo.fields['PCR2000'].data)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Using PANDAS to get that R Feel"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pandas as PANDAS\n",
      "ids = [ssdo.order2Master[i] for i in range(ssdo.numObs)]\n",
      "convertDictDF = {}\n",
      "for fieldName, fieldObject in ssdo.fields.items():\n",
      "    convertDictDF[fieldName] = fieldObject.data\n",
      "df = PANDAS.DataFrame(convertDictDF, index = ids)\n",
      "print(df)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Advanced Analysis [SciPy Example - KMeans]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as NUM\n",
      "import scipy.cluster.vq as CLUST\n",
      "import arcgisscripting as ARC\n",
      "X = NUM.empty((ssdo.numObs,2), dtype = float)\n",
      "X[:,0] = df['PCR2000']\n",
      "X[:,1] = df['POP2000']\n",
      "whiteData = CLUST.whiten(X)\n",
      "centers, distortion = CLUST.kmeans(whiteData, 5)\n",
      "groups = ARC._ss.closest_centroid(whiteData, centers)\n",
      "print(groups)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ARCPY.env.overwriteOutput = True\n",
      "outputFC = r'C:\\Data\\UC\\DevSummit15\\OpenSource\\output\\kmeans_ca.shp'\n",
      "outField = SSDO.CandidateField('SS_GROUP', 'LONG', groups)\n",
      "outFields = {'SS_GROUP': outField}\n",
      "ssdo.output2NewFC(outputFC, outFields, appendFields = [\"NEW_NAME\", \"PERCNOHS\"])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}