Notebooks/SampleDesign.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jshogland/SpatialModelingTutorials/blob/main/Notebooks/SampleDesign.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "28fd79a1-e56f-457c-8c15-a94ee0f65231",
      "metadata": {
        "id": "28fd79a1-e56f-457c-8c15-a94ee0f65231"
      },
      "source": [
        "# Sample Design (Microsoft Planet Computer) \n",
        "## This notebook demonstrates how to create a stratified sample design that is spread in predictor variable space and that can be used to build a machine learning model.\n",
        "### Author: John Hogland 3/21/2023"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "64bc00aa-36d8-4ff4-a7ce-4a513079141a",
      "metadata": {
        "id": "64bc00aa-36d8-4ff4-a7ce-4a513079141a"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "9e1054ab-f793-41e2-8b1e-4f2c055ca2ff",
      "metadata": {
        "tags": [],
        "id": "9e1054ab-f793-41e2-8b1e-4f2c055ca2ff"
      },
      "source": [
        "## Install packages"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "639dea01-c04c-47e2-9677-99e5c193d107",
      "metadata": {
        "id": "639dea01-c04c-47e2-9677-99e5c193d107"
      },
      "outputs": [],
      "source": [
        "!pip install --upgrade gdown\n",
        "!pip install --upgrade numba\n",
        "!pip install --upgrade geopandas\n",
        "!pip install mapclassify\n",
        "!pip install --upgrade datascience\n",
        "!pip install --upgrade gym\n",
        "!pip install --upgrade folium\n",
        "!pip install plotly\n",
        "!pip install laspy[lazrs,laszip]\n",
        "!pip install raster_tools\n",
        "!pip install stackstac\n",
        "!pip install planetary_computer\n",
        "!pip install pystac_client\n",
        "!pip install leafmap xarray_leaflet\n",
        "!pip install localtileserver"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "14796e91-e08c-4faa-ba5a-537e330f7ec4",
      "metadata": {
        "id": "14796e91-e08c-4faa-ba5a-537e330f7ec4"
      },
      "source": [
        "## Get supporting python file"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "47464719-6be8-4fcb-b83b-38f2dc0961d6",
      "metadata": {
        "id": "47464719-6be8-4fcb-b83b-38f2dc0961d6"
      },
      "outputs": [],
      "source": [
        "import gdown, zipfile\n",
        "\n",
        "url='https://drive.google.com/file/d/1dy7bnPKc4BPvHlH-PkrObXwb-SW9nv7n/view?usp=sharing'\n",
        "outfl= r'./rs_las.py'\n",
        "gdown.download(url=url,output=outfl,quiet=False,fuzzy=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "9750cc2c-9fc0-4312-8e1f-6d856bc9e3d7",
      "metadata": {
        "id": "9750cc2c-9fc0-4312-8e1f-6d856bc9e3d7"
      },
      "source": [
        "## The Process\n",
        "### In this notebook we will be using derivatives from Lidar data acquired from Microsoft's Planetary Computer to build a stratified random sample spread across predictor variable space. This notebooks builds upon the Lidar [Download](https://github.com/jshogland/SpatialModelingTutorials/blob/main/Notebooks/LidarDownload.ipynb) and [Processing](https://github.com/jshogland/SpatialModelingTutorials/blob/main/Notebooks/LidarProcessing.ipynb) notebooks at [Spatial Modeling Tutorials](https://github.com/jshogland/SpatialModelingTutorials).  \n"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "971acc1d-9dee-4b02-bffb-50aaa3352a43",
      "metadata": {
        "tags": [],
        "id": "971acc1d-9dee-4b02-bffb-50aaa3352a43"
      },
      "source": [
        "### Steps\n",
        "1. Import python libraries and download laz files\n",
        "2. Create Point Cloud raster surface predictors \n",
        "3. Visualize predictor surfaces\n",
        "4. Create a large random sample and extract predictor values\n",
        "5. Create a stratified random sample of 150 locations that are well spread\n",
        "6. Collect field data for those locations"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "8695de31-9d43-4d74-a2da-4941328c50fc",
      "metadata": {
        "tags": [],
        "id": "8695de31-9d43-4d74-a2da-4941328c50fc"
      },
      "source": [
        "## Step 1: Import libraries and get data\n",
        "### Import libary"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "77411e51-f774-43be-8c10-ab53eec14b6b",
      "metadata": {
        "tags": [],
        "id": "77411e51-f774-43be-8c10-ab53eec14b6b"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import geopandas as gpd\n",
        "import pandas as pd\n",
        "import rs_las\n",
        "\n",
        "import planetary_computer\n",
        "import pystac_client\n",
        "import requests\n",
        "import stackstac\n",
        "\n",
        "from dask.diagnostics import ProgressBar\n",
        "\n",
        "\n",
        "from raster_tools import Raster\n",
        "from raster_tools import zonal\n",
        "\n",
        "from shapely.geometry import shape\n",
        "from shapely.ops import transform\n",
        "\n",
        "import leafmap.leafmap as leafmap\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "5fd8e9db-e4c0-496a-8d9e-b713206583a7",
      "metadata": {
        "id": "5fd8e9db-e4c0-496a-8d9e-b713206583a7"
      },
      "source": [
        "## *If you already have the laz files for Helena from the [Lidar Download](https://github.com/jshogland/SpatialModelingTutorials/blob/main/Notebooks/LidarDownload.ipynb) notebook, you can skip to Step 2*\n",
        "### Set the point location to Helena MT"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "06e3747f-20e2-449f-8330-4f3108dbef39",
      "metadata": {
        "tags": [],
        "id": "06e3747f-20e2-449f-8330-4f3108dbef39"
      },
      "outputs": [],
      "source": [
        "loc = {\"type\": \"Point\", \"coordinates\": [-112.027, 46.64]} # just a little north of town\n",
        "geom = shape(loc) # create a shapely point\n",
        "geom_buff=geom.buffer(0.01) #buffer area to create a area of interest (AOI)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "5752e49b-704d-42ba-8794-98a9e979715e",
      "metadata": {
        "id": "5752e49b-704d-42ba-8794-98a9e979715e"
      },
      "source": [
        "### Get the laz data from Planetary Computer\n",
        "#### We will be using laz files downloaded from Planetary Computer for a small area around Helena MT. To download the laz files run the following cell or work through the [LidarDownload.ipynb](https://github.com/jshogland/SpatialModelingTutorials/blob/main/Notebooks/LidarDownload.ipynb) notebook."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "80da7034-0a64-4bf9-a80d-0c6a473f0651",
      "metadata": {
        "tags": [],
        "id": "80da7034-0a64-4bf9-a80d-0c6a473f0651"
      },
      "outputs": [],
      "source": [
        "catalog = pystac_client.Client.open(\"https://planetarycomputer.microsoft.com/api/stac/v1\", modifier=planetary_computer.sign_inplace)\n",
        "\n",
        "search = catalog.search(collections=[\"3dep-lidar-copc\"], intersects=geom_buff)\n",
        "ic = search.get_all_items()\n",
        "\n",
        "#create a output folder if it does not already exist\n",
        "\n",
        "output_folder= './lidar_cloud'\n",
        "if(not os.path.exists(output_folder)): os.mkdir(output_folder)\n",
        "\n",
        "for tile in ic:\n",
        "        url = tile.assets[\"data\"].href\n",
        "        nm = tile.id + '.copc.laz'\n",
        "        print('downloading ' + nm)\n",
        "        r=requests.get(url,allow_redirects=True)\n",
        "        open(output_folder+'/'+nm,'wb').write(r.content)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "1e428c45-ae87-479f-9ae3-ab8fabc88f2b",
      "metadata": {
        "tags": [],
        "id": "1e428c45-ae87-479f-9ae3-ab8fabc88f2b"
      },
      "source": [
        "## *If you have already saved predictor surfaces for Helena from the [Lidar Processing notebook](https://github.com/jshogland/SpatialModelingTutorials/blob/main/Notebooks/LidarProcessing.ipynb), you can skip to Step 3*\n",
        "## Step 2: Create Laz Predictors Surfaces (Raster objects 30m)\n",
        "\n",
        "- zmean\n",
        "- zmax\n",
        "- zp10, zp25, zp50, zp75, zp90, zp99\n",
        "- zfcc\n",
        "- zfsc\n",
        "- zcv\n",
        "\n",
        "### Use [rs_las](./rs_las.py) to create the metrics (aggtypes) and save to a conventional raster type"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "bb6a40bf-a596-41bc-b5e3-8d6f88dde960",
      "metadata": {
        "tags": [],
        "id": "bb6a40bf-a596-41bc-b5e3-8d6f88dde960"
      },
      "outputs": [],
      "source": [
        "%%time\n",
        "laz_dir = './lidar_cloud'\n",
        "laz_files_lst=[]\n",
        "for f in os.listdir(laz_dir):\n",
        "    if f[-4:]=='.laz': laz_files_lst.append(laz_dir+\"/\"+f)\n",
        "\n",
        "res=30\n",
        "\n",
        "gdf=rs_las.build_extents(laz_files_lst)\n",
        "\n",
        "aggtypes=['mean','max','quantile:0.1,0.25,0.5,0.75,0.9,0.99','zfcc','zfsc','zcv']\n",
        "rs_z=rs_las.get_raster(gdf,res,aggtypes)\n",
        "rs_z.save('Lidar_30_metrics.tif')"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "cb9445c8-4a6b-4c45-b869-d0c2c252fa47",
      "metadata": {
        "id": "cb9445c8-4a6b-4c45-b869-d0c2c252fa47"
      },
      "source": [
        "## Step 3: Visualize predictor surfaces using [Leafmap](https://leafmap.org/leafmap/)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "e21a041f-3b40-4e04-a936-0794f52586b8",
      "metadata": {
        "tags": [],
        "id": "e21a041f-3b40-4e04-a936-0794f52586b8"
      },
      "outputs": [],
      "source": [
        "#Specify the name of the predictor raster stack\n",
        "outname='Lidar_30_metrics.tif'\n",
        "\n",
        "#Create a leafmap map object\n",
        "m = leafmap.Map()\n",
        "\n",
        "#Add ESRI's World imagery as a tile layer\n",
        "m.add_tile_layer(url='https://server.arcgisonline.com/ArcGIS/rest/services/World_Imagery/MapServer/tile/{z}/{y}/{x}', name = 'Esri Imagery',attribution='ESRI')\n",
        "\n",
        "#Add each band of the raster to the map object\n",
        "m.add_raster(outname,bands=[11], layer_name='Coefficent of Variation (z)')\n",
        "m.add_raster(outname,bands=[10], layer_name='% Shrub Cover')\n",
        "m.add_raster(outname,bands=[9], layer_name='% Canopy Cover')\n",
        "m.add_raster(outname,bands=[8], layer_name='99% (z)')\n",
        "m.add_raster(outname,bands=[7], layer_name='90% (z)')\n",
        "m.add_raster(outname,bands=[6], layer_name='75% (z)')\n",
        "m.add_raster(outname,bands=[5], layer_name='50% (z)')\n",
        "m.add_raster(outname,bands=[4], layer_name='25% (z)')\n",
        "m.add_raster(outname,bands=[3], layer_name='10% (z)')\n",
        "m.add_raster(outname,bands=[2], layer_name='max (z)')\n",
        "m.add_raster(outname,bands=[1], layer_name='mean (z)')\n",
        "\n",
        "#show the map\n",
        "m"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "6a186999-85a4-44ed-98c6-f3280780d2e5",
      "metadata": {
        "id": "6a186999-85a4-44ed-98c6-f3280780d2e5"
      },
      "source": [
        "## Step 4: Create large random sample \n",
        "- create a large random sample of spatial locations (~100 times the sample size)\n",
        "- extract the predictor surface values for each location and populate point values\n",
        "- we will use this sample to select a well spread and balanced sample using kmean clustering"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "1c1a47ca-02ad-42d1-842d-422acfafe71b",
      "metadata": {
        "tags": [],
        "id": "1c1a47ca-02ad-42d1-842d-422acfafe71b"
      },
      "outputs": [],
      "source": [
        "#create large random sample\n",
        "bn=150*100\n",
        "predictors84=Raster(outname)\n",
        "laz_dir = './lidar_cloud'\n",
        "laz_files_lst=[]\n",
        "for f in os.listdir(laz_dir):\n",
        "    if f[-4:]=='.laz': laz_files_lst.append(laz_dir+\"/\"+f)\n",
        "\n",
        "gdf=rs_las.build_extents(laz_files_lst)\n",
        "\n",
        "\n",
        "bnds=gdf.total_bounds\n",
        "\n",
        "pred=Raster(predictors84.xdata.rio.reproject(gdf.crs))\n",
        "\n",
        "x=np.random.randint(bnds[0],bnds[2],bn)\n",
        "y=np.random.randint(bnds[1],bnds[3],bn)\n",
        "b_loc=gpd.GeoDataFrame(geometry=gpd.points_from_xy(x,y,crs=gdf.crs))\n",
        "b_loc2=b_loc[b_loc.intersects(gdf.unary_union)]\n",
        "\n",
        "p1=gdf.plot(facecolor=\"none\",edgecolor='orange',figsize=(15,15),linewidth=2,zorder=1)\n",
        "p2=b_loc2.plot(ax=p1,color='blue',zorder=0)\n",
        "p2"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "3e64206d-b79f-43e3-af33-a8a044c10c71",
      "metadata": {
        "tags": [],
        "id": "3e64206d-b79f-43e3-af33-a8a044c10c71"
      },
      "outputs": [],
      "source": [
        "#extract predictor surface values\n",
        "pvls=zonal.extract_points_eager(b_loc2,pred,'Lidar').compute()\n",
        "\n",
        "#create common index for pvls and locations\n",
        "pvls['pltid']=list(b_loc2.index.values)*11\n",
        "pvls['band']=pvls['band'].astype(str)\n",
        "b_loc2['pltid']=b_loc2.index.values\n",
        "\n",
        "#transpose pvls, fill NaN values, and merge location with surface values and visualize\n",
        "b_sample=b_loc2.merge(pvls.pivot(index='pltid',columns='band',values='Lidar').fillna(0),on='pltid')\n",
        "b_sample"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "04744e54-3fd9-413e-ba7d-bad2f648021a",
      "metadata": {
        "id": "04744e54-3fd9-413e-ba7d-bad2f648021a"
      },
      "source": [
        "## Step 5: Create a stratified random sample of ~150 locations that is well spread.\n",
        "### We are sampling locations because it is too expensive to completely inventory a natural resource (e.g., Tree Basal Area). From the sample we can make estimates about the population, i.e. the study area (design based inference). However, we can also use this sample to train a model that can be used to gain insights to smaller spatial domains (the cell). Using our sample we will create a model that can be used to predict estimates using the inventory of Lidar data based metrics ([LidarEstimatingBaa.ipynb]()) and use those estimates to infer population statistics (model based inference). To build a stratified random sample we will be using information from our Lidar based surfaces. \n",
        "- we will subset our large random sample to create a well spread sample\n",
        "- we will be using sklearn's k-means algorithm to create 10 clusters\n",
        "- from each cluster we will select ~15 random locations\n",
        "- finally let's visualize the spatial location of each field observation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "a918128d-d010-47c3-b5b2-78896916196e",
      "metadata": {
        "tags": [],
        "id": "a918128d-d010-47c3-b5b2-78896916196e"
      },
      "outputs": [],
      "source": [
        "from sklearn.cluster import KMeans\n",
        "c=10\n",
        "n=15\n",
        "X = b_sample[['1','2','3','4','5','6','7','8','9','10','11']]\n",
        "mdl=KMeans(n_clusters=10).fit(X)\n",
        "b_sample['cluster']=(mdl.fit_predict(X)).astype('str')\n",
        "test=np.unique(b_sample.cluster,return_counts=True)\n",
        "n=int(c/np.sum(test[1]>=n) * n)\n",
        "\n",
        "def samp_c(x,size):\n",
        "    if(x.shape[0]>size):return x.sample(size)\n",
        "\n",
        "wsb_sample=b_sample.groupby('cluster', group_keys=False).apply(lambda x: samp_c(x,n)) \n",
        "print(\"The total sample size = \" + str(wsb_sample.shape[0]))\n",
        "\n",
        "p1=gdf.plot(facecolor=\"none\",edgecolor='orange',figsize=(10,10),linewidth=2,zorder=1)\n",
        "p2=wsb_sample.plot(ax=p1,column='cluster',legend=True,zorder=0)\n",
        "p2"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Exercise 1: How different are mean values?\n",
        "- quantify the difference in mean lidar metrics between the large random sample and the smaller stratified random sample.\n",
        "- which mean is correct?\n",
        "- what needs to happen to the stratified random sample to give population based estimates?\n",
        "- what would happen if we increased the number of clusters?"
      ],
      "metadata": {
        "id": "ClNE_yRrTDWw"
      },
      "id": "ClNE_yRrTDWw"
    },
    {
      "cell_type": "markdown",
      "id": "89394463-12eb-48e6-8c39-43ef3ad56ae1",
      "metadata": {
        "id": "89394463-12eb-48e6-8c39-43ef3ad56ae1"
      },
      "source": [
        "## Step 6: Collect field data\n",
        "- unfortunately we can't go to the field in this exercise\n",
        "- instead, we will create a relationship between our Lidar metrics and basal area as a surrogate for actual field data collection\n",
        "- our measurement we will be \"collecting\" for each location (field plot) is total tree basal area per acre (BAA) measured in ft<sup>2</sup> per acre\n",
        "- the equation to calculate our wsb_sample is $ BAA = 20 + 2*zcv + 1.5*zfcc + 0.9*zfsc + 0.53*zp90 + random()*10) $\n",
        "- save the dataframe as a shape file"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "2091b6f1-edad-4b18-b9f9-bd7250aa9d7c",
      "metadata": {
        "id": "2091b6f1-edad-4b18-b9f9-bd7250aa9d7c"
      },
      "outputs": [],
      "source": [
        "#create BAA and visualize\n",
        "wsb_sample['BAA']=20 + (2*wsb_sample['11']) + (1.5*wsb_sample['9']) + (0.9*wsb_sample['10']) + (0.530*wsb_sample['7']) + (np.random.random(wsb_sample.shape[0])*10)\n",
        "wsb_sample['BAA'].plot.hist()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "c9ad14e5-5e67-45cd-92be-5e20fe083096",
      "metadata": {
        "tags": [],
        "id": "c9ad14e5-5e67-45cd-92be-5e20fe083096"
      },
      "outputs": [],
      "source": [
        "#save the dataframe\n",
        "wsb_sample.to_file('train.shp')"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "0c80290a-ad06-4c61-88e3-b11fbe2e7b67",
      "metadata": {
        "id": "0c80290a-ad06-4c61-88e3-b11fbe2e7b67"
      },
      "source": [
        "# This ends the Sample Design notebook\n",
        "## Check out the other notebooks:\n",
        "- https://github.com/jshogland/SpatialModelingTutorials/tree/main/Notebooks\n",
        "## References\n",
        "- Spatial Modeling Notebooks:  https://github.com/jshogland/SpatialModelingTutorials\n",
        "- Raster-Tools GitHub: https://github.com/UM-RMRS/raster_tools\n",
        "- Hogland's Spatial Solutions: https://sites.google.com/view/hoglandsspatialsolutions/home\n",
        "- Dask: https://dask.org/\n",
        "- Geopandas:https://geopandas.org/en/stable/\n",
        "- Xarray: https://docs.xarray.dev/en/stable/\n",
        "- Jupyter: https://jupyter.org/\n",
        "- Anaconda:https://www.anaconda.com/\n",
        "- VS Code: https://code.visualstudio.com/\n",
        "- ipywidgets: https://ipywidgets.readthedocs.io/en/latest/\n",
        "- numpy:https://numpy.org/\n",
        "- matplotlib:https://matplotlib.org/\n",
        "- folium: https://python-visualization.github.io/folium/\n",
        "- pandas: https://pandas.pydata.org/\n",
        "- sklearn: https://scikit-learn.org/stable/index.html"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "RStools Python 3.10",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.10.9"
    },
    "colab": {
      "provenance": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}