Add indicator percentpop euclidean proximity to openspace

city_metrix/layers/layer.py

@@ -135,6 +135,9 @@ def mean(self):
 
     def count(self):
         return self._zonal_stats("count")
+
+    def sum(self):
+        return self._zonal_stats("sum")
 
     def _zonal_stats(self, stats_func):
         if box(*self.zones.total_bounds).area <= MAX_TILE_SIZE_DEGREES**2:
@@ -315,6 +318,8 @@ def _aggregate_stats(df, stats_func):
     elif stats_func == "mean":
         # mean must weight by number of pixels used for each tile
         return (df["mean"] * df["count"]).sum() / df["count"].sum()
+    elif stats_func == "sum":
+        return df["sum"].sum()
 
 
 def get_stats_funcs(stats_func):

city_metrix/layers/open_street_map.py

@@ -35,9 +35,16 @@ class OpenStreetMapClass(Enum):
 
 
 class OpenStreetMap(Layer):
-    def __init__(self, osm_class=None, **kwargs):
+    """
+    Attributes:
+        osm_class: Enum value from OpenStreetMapClass
+        buffer_distance: meters distance for buffer around osm features
+    """
+
+    def __init__(self, osm_class=None, buffer_distance=None, **kwargs):
         super().__init__(**kwargs)
         self.osm_class = osm_class
+        self.buffer_distance = buffer_distance  # meters
 
     def get_data(self, bbox):
         north, south, east, west = bbox[3], bbox[1], bbox[0], bbox[2]
@@ -74,4 +81,7 @@ def get_data(self, bbox):
         crs = get_utm_zone_epsg(bbox)
         osm_feature = osm_feature.to_crs(crs)
 
+        if self.buffer_distance:
+            osm_feature['geometry'] = osm_feature.geometry.buffer(self.buffer_distance)
+
         return osm_feature

city_metrix/metrics/__init__.py

@@ -1,7 +1,8 @@
 from .built_land_without_tree_cover import built_land_without_tree_cover
 from .built_land_with_low_surface_reflectivity import built_land_with_low_surface_reflectivity
 from .built_land_with_high_land_surface_temperature import built_land_with_high_land_surface_temperature
+from .era_5_met_preprocessing import era_5_met_preprocessing
 from .mean_tree_cover import mean_tree_cover
-from .urban_open_space import urban_open_space
 from .natural_areas import natural_areas
-from .era_5_met_preprocessing import era_5_met_preprocessing
+from .pop_open_space import pop_open_space
+from .urban_open_space import urban_open_space

city_metrix/metrics/pop_open_space.py

@@ -0,0 +1,13 @@
+from geopandas import GeoDataFrame, GeoSeries
+from city_metrix.layers import OpenStreetMap, OpenStreetMapClass, WorldPop
+
+
+def pop_open_space(zones: GeoDataFrame, buffer_distance=400) -> GeoSeries:
+# (Later add agesex_classes)
+    pop = WorldPop()
+    open_space = OpenStreetMap(osm_class=OpenStreetMapClass.OPEN_SPACE, buffer_distance=buffer_distance)
+
+    pop_open_space_sum = pop.mask(open_space).groupby(zones).sum()
+    pop_sum = pop.groupby(zones).sum()
+
+    return pop_open_space_sum / pop_sum

tests/test_metrics.py

@@ -44,6 +44,13 @@ def test_natural_areas():
     assert expected_zone_size == actual_indicator_size
 
 
+def test_pop_open_space():
+    indicator = pop_open_space(ZONES)
+    expected_zone_size = ZONES.geometry.size
+    actual_indicator_size = indicator.size
+    assert expected_zone_size == actual_indicator_size
+
+
 def test_urban_open_space():
     indicator = urban_open_space(ZONES)
     expected_zone_size = ZONES.geometry.size