chore: start adding markdown to new notebook

kraina-ai · Jan 24, 2025 · f11e025 · f11e025
1 parent fa9bb62
commit f11e025
Showing 1 changed file with 24 additions and 30 deletions.
diff --git a/examples/use_cases/simple_ml_with_overture_maps_data.ipynb b/examples/use_cases/simple_ml_with_overture_maps_data.ipynb
@@ -1,5 +1,16 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Simple ML model using Overture Maps data\n",
+    "\n",
+    "Prerequisites:\n",
+    "- 12 GB of RAM\n",
+    "- Installed libraries: `srai[osm,overturemaps,plotting]`, `contextily`, `seaborn`, `xgboost`, `pypalettes`"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -18,15 +29,13 @@
     "from h3 import int_to_str, str_to_int\n",
     "from h3ronpy import grid_disk_aggregate_k\n",
     "from pypalettes import load_cmap\n",
-    "from pytorch_lightning import seed_everything\n",
     "from sklearn.preprocessing import StandardScaler\n",
     "\n",
     "from srai.embedders import ContextualCountEmbedder\n",
-    "\n",
-    "# from srai.embedders import CountEmbedder\n",
     "from srai.h3 import h3_to_shapely_geometry\n",
     "from srai.joiners import IntersectionJoiner\n",
     "from srai.loaders import OSMOnlineLoader\n",
+    "from srai.loaders.overturemaps_loader import OvertureMapsLoader\n",
     "from srai.neighbourhoods import H3Neighbourhood\n",
     "from srai.regionalizers import H3Regionalizer, geocode_to_region_gdf"
    ]
@@ -37,8 +46,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "SEED = 71\n",
-    "seed_everything(SEED)"
+    "SEED = 71"
    ]
   },
   {
@@ -163,7 +171,7 @@
     ")\n",
     "h3_regions[\n",
     "    (h3_regions.city == \"Madrid\") & (h3_regions.distance_to_station > H3_PREDICTION_RANGE)\n",
-    "].plot(ax=ax, color=\"gray\", alpha=0.2)\n",
+    "].plot(ax=ax, color=\"gray\", alpha=0.3)\n",
     "bicycle_stations_per_city[\"Madrid\"].representative_point().plot(ax=ax, color=\"black\", markersize=1)\n",
     "\n",
     "cx.add_basemap(ax, crs=h3_regions.crs, source=cx.providers.CartoDB.PositronNoLabels, zoom=13)\n",
@@ -178,12 +186,9 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from srai.loaders.overturemaps_loader import OvertureMapsLoader\n",
-    "\n",
     "OVERTURE_MAPS_HIERARCHY_DEPTH_VALUES = {\n",
     "    (\"base\", \"infrastructure\"): 1,\n",
     "    (\"base\", \"land\"): 1,\n",
-    "    # (\"base\", \"land_cover\"): 1,\n",
     "    (\"base\", \"land_use\"): 1,\n",
     "    (\"base\", \"water\"): 1,\n",
     "    (\"transportation\", \"segment\"): 2,\n",
@@ -197,20 +202,16 @@
     "    hierarchy_depth=list(OVERTURE_MAPS_HIERARCHY_DEPTH_VALUES.values()),\n",
     "    include_all_possible_columns=False,\n",
     ").load(area=h3_regions)\n",
-    "features  # 637785 rows × 111 columns\n",
+    "features\n",
     "\n",
+    "# if you want to use OpenStreetMap data instead you can use `OSMPbfLoader`\n",
+    "# with `GEOFABRIK_LAYERS` filter\n",
+    "\n",
+    "# from srai.loaders.osm_loaders.filters import GEOFABRIK_LAYERS\n",
+    "# from srai.loaders.osm_loaders import OSMPbfLoader\n",
     "# features = OSMPbfLoader().load(area=h3_regions, tags=GEOFABRIK_LAYERS)"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "features.columns.to_list()"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -233,16 +234,10 @@
     "    concatenate_vectors=False,\n",
     "    count_subcategories=False,\n",
     ").transform(regions_gdf=h3_regions, features_gdf=features, joint_gdf=joint)\n",
-    "embeddings"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "embeddings = embeddings.drop(columns=[c for c in embeddings.columns if \"bicycle_rental\" in c])\n",
+    "\n",
+    "# If you are using OpenStreetMap data, remeber to remove bicycle_sharing stations from the dataset\n",
+    "# embeddings = embeddings.drop(columns=[c for c in embeddings.columns if \"bicycle_rental\" in c])\n",
+    "\n",
     "embeddings"
    ]
   },
@@ -462,7 +457,6 @@
     "            \"pad\": -0.05,\n",
     "        },\n",
     "        vmin=max(0, city_data[\"predicted_distance_to_station\"].min()),\n",
-    "        # vmax=max(H3_PREDICTION_MAX_DISTANCE, city_data[\"predicted_distance_to_station\"].max()),\n",
     "        vmax=city_data[\"predicted_distance_to_station\"].max(),\n",
     "    )\n",
     "    bicycle_stations_per_city[city_name].representative_point().plot(\n",