reprs for H3Shape objects should work better with geopandas

uber · Nov 12, 2023 · d60a5af · d60a5af
1 parent d52679a
commit d60a5af
Showing 1 changed file with 111 additions and 5 deletions.
diff --git a/poly/plotting.ipynb b/poly/plotting.ipynb
@@ -1,5 +1,34 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "aa3a6a1e-b947-4e92-a5e4-c3dac6621885",
+   "metadata": {},
+   "source": [
+    "# notes\n",
+    "\n",
+    "## projections\n",
+    "\n",
+    "- 4326 # wgs84 (for H3 compatibility)\n",
+    "    - H3 **does not use** wgs84/epsg4326, but it **does use** the \"authalic radius\" from that standard\n",
+    "- 3857 # web mercator (for plotting)\n",
+    "- 2263 # new york (just what's used in the example below)\n",
+    "\n",
+    "## geoviews\n",
+    "\n",
+    "- tile options: https://github.com/holoviz/geoviews/blob/main/geoviews/tile_sources.py\n",
+    "    - also this, but i don't see a good way to programatically get all the options: https://geoviews.org/gallery/matplotlib/tile_sources.html \n",
+    "- active tools: https://stackoverflow.com/questions/57767169/set-box-zoom-or-pan-or-wheel-zoom-as-default-in-using-holoviews-or-hvplot\n",
+    "- bokeh pallets: https://docs.bokeh.org/en/latest/docs/reference/palettes.html\n",
+    "- Bokeh \"tap\" tool seems to do something on this website, but i can't get the same functionality (isolating a single geometry) locally: https://geoviews.org/gallery/bokeh/new_york_boroughs.html\n",
+    "\n",
+    "## folium\n",
+    "\n",
+    "- seems to be most popular, but they make you manually determine the center and zoom level!\n",
+    "    - too much work. geoviews does it automatically\n",
+    "    - https://geopandas.org/en/stable/gallery/polygon_plotting_with_folium.html"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -20,11 +49,7 @@
    "id": "9e971ca2-24a4-43ba-8a90-83f6091b462b",
    "metadata": {},
    "outputs": [],
-   "source": [
-    "2263 # new york\n",
-    "4326 # wgs84\n",
-    "3857 # web mercator"
-   ]
+   "source": []
   },
   {
    "cell_type": "code",
@@ -147,12 +172,93 @@
     "bar(df, 3857)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "c6a0ea52-0721-496e-bb1f-50faebd38493",
+   "metadata": {},
+   "source": [
+    "# converting to cells\n",
+    "\n",
+    "- example of a common bug users will run into"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c701a9b6-a26d-44b7-beb1-9981d9ea22d3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df = geopandas.read_file(geodatasets.get_path('nybb'))\n",
+    "df.crs"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ae6bec98-1a7f-4223-8990-04a3169c2911",
+   "metadata": {},
+   "source": [
+    "The following code works, but runs for a while and then breaks:\n",
+    "\n",
+    "```python\n",
+    "df.geometry.apply(lambda x: h3.geo_to_cells(x, 9))\n",
+    "```\n",
+    "\n",
+    "The bug is that we first need to convert to the WGS84 authalic radius system:\n",
+    "`df = df.to_crs(epsg=4326)`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c432273e-d03f-4a29-8788-b869f1b9cf9b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#df.geometry.apply(lambda x: h3.geo_to_cells(x, 9))\n",
+    "df"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
    "id": "831de887-a746-430b-a586-5e7069581d95",
    "metadata": {},
    "outputs": [],
+   "source": [
+    "df84 = df.to_crs(epsg=4326)\n",
+    "df84"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "65f2c481-2b80-4acd-932d-5c873e58cccd",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# TODO: this needs to show H3MultiPoly\n",
+    "out = df84.geometry.apply(lambda x: h3.geo_to_cells(x, 9)).apply(lambda x: h3.cells_to_h3shape(x))\n",
+    "out"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "adacc9ea-3370-4060-82ff-8a10321e7dc1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for a in out:\n",
+    "    print(a)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "aac772f7-255f-4a61-954f-ac04240f163d",
+   "metadata": {},
+   "outputs": [],
    "source": []
   }
  ],