Merge pull request #392 from oturns/sedafix

add isochrone and silhouette tests

geosnap/tests/test_diagnostics.py

@@ -1,17 +1,17 @@
 from geosnap.io import get_census
 from geosnap import DataStore
-from geosnap.analyze import find_k, find_region_k
+from geosnap.analyze import find_k, find_region_k, cluster, regionalize
 from numpy.testing import assert_array_equal, assert_array_almost_equal
 
+reno = get_census(msa_fips="39900", datastore=DataStore(), years=[2010])
+columns = [
+    "median_household_income",
+    "p_poverty_rate",
+    "p_unemployment_rate",
+]
 
-def test_find_k():
-    reno = get_census(msa_fips="39900", datastore=DataStore(), years=[2010])
-    columns = [
-        "median_household_income",
-        "p_poverty_rate",
-        "p_unemployment_rate",
-    ]
 
+def test_find_k():
     ks = find_k(
         reno,
         columns=columns,
@@ -20,17 +20,10 @@ def test_find_k():
     )
     # Aspatial Clusters
 
-    assert_array_almost_equal( ks.T.values[0], [2,2,2])
-
-def test_find_region_k():
-
-    reno = get_census(msa_fips="39900", datastore=DataStore(), years=[2010])
-    columns = [
-        "median_household_income",
-        "p_poverty_rate",
-        "p_unemployment_rate",
-    ]
+    assert_array_almost_equal(ks.T.values[0], [2, 2, 2])
 
+
+def test_find_region_k():
     ks = find_region_k(
         reno,
         columns=columns,
@@ -39,4 +32,24 @@ def test_find_region_k():
     )
     # Aspatial Clusters
 
-    assert_array_almost_equal( ks.values[0],[2.,2.,2.,2.,2.])
+    assert_array_almost_equal(ks.values[0], [2.0, 2.0, 2.0, 2.0, 2.0])
+
+
+def test_cluster_diagnostics():
+    ward, ward_mod = cluster(
+        reno, columns=columns, method="ward", n_clusters=5, return_model=True
+    )
+    assert ward_mod.silhouette_score.round(4) == 0.2991
+    assert ward_mod.davies_bouldin_score.round(4) == 1.0336
+    assert ward_mod.calinski_harabasz_score.round(4) == 88.6627
+
+
+def test_region_diagnostics():
+    ward, ward_mod = regionalize(
+        reno, columns=columns, method="ward_spatial", n_clusters=5, return_model=True
+    )
+    assert ward_mod[2010].boundary_silhouette.boundary_silhouette.mean().round(4) ==  0.2076
+    assert ward_mod[2010].path_silhouette.path_silhouette.mean().round(4) ==  -0.0801
+    assert ward_mod[2010].silhouette_scores.silhouette_score.mean().round(4) ==0.063
+    assert ward_mod[2010].nearest_label.nearest_label.head().tolist() == [1, 2, 0, 2, 4]
+

geosnap/tests/test_isochrones.py

@@ -0,0 +1,42 @@
+from geosnap.analyze import (
+    isochrones_from_id,
+    isochrones_from_gdf,
+)
+from geosnap.io import get_acs
+from geosnap import DataStore
+
+import pandana as pdna
+import geopandas as gpd
+import os
+
+if not os.path.exists("./41740.h5"):
+    import quilt3 as q3
+
+    b = q3.Bucket("s3://spatial-ucr")
+    b.fetch("osm/metro_networks_8k/41740.h5", "./41740.h5")
+
+datasets = DataStore()
+sd_tracts = get_acs(datasets, county_fips="06073", years=[2018])
+sd_network = pdna.Network.from_hdf5("41740.h5")
+example_origin = 1985327805
+
+
+def test_isos_from_ids():
+    iso = isochrones_from_id(example_origin, sd_network, threshold=1600)
+    assert iso.area.round(6).tolist()[0] == 0.000128
+
+
+def test_isos_from_gdf():
+    sd_network.nodes_df["geometry"] = gpd.points_from_xy(
+        sd_network.nodes_df.x, sd_network.nodes_df.y
+    )
+    example_point = gpd.GeoDataFrame(
+        sd_network.nodes_df.loc[example_origin]
+    ).T.set_geometry("geometry")
+    example_point = example_point.set_crs(4326)
+    t = isochrones_from_gdf(
+        origins=example_point,
+        network=sd_network,
+        threshold=1600,
+    )
+    assert t.area.round(8).tolist()[0] == 0.00012821