momepy.percentile#

momepy.percentile(y, graph, q=[25, 50, 75])[source]#

Calculates linearly weighted percentiles of y values using the neighbourhoods and weights defined in graph.

The specific interpolation method implemented is “hazen”.

Parameters:

ySeries: A Series containing the values to be analysed.
graphlibpysal.graph.Graph: A spatial weights matrix for the data.
qarray-like (default [25, 50, 75]): The percentiles to return.

Returns:

Dataframe: A Dataframe with columns as the results for each percentile

Examples

>>> from libpysal import graph
>>> path = momepy.datasets.get_path("bubenec")
>>> buildings = geopandas.read_file(path, layer="buildings")
>>> buildings.head()
   uID                                           geometry
0    1  POLYGON ((1603599.221 6464369.816, 1603602.984...
1    2  POLYGON ((1603042.88 6464261.498, 1603038.961 ...
2    3  POLYGON ((1603044.65 6464178.035, 1603049.192 ...
3    4  POLYGON ((1603036.557 6464141.467, 1603036.969...
4    5  POLYGON ((1603082.387 6464142.022, 1603081.574...

Define spatial graph:

>>> knn5 = graph.Graph.build_knn(buildings.centroid, k=5)
>>> knn5
<Graph of 144 nodes and 720 nonzero edges indexed by
 [0, 1, 2, 3, 4, ...]>

Percentiles of building area within 5 nearest neighbors:

>>> momepy.percentile(buildings.area, knn5).head()
               25          50           75
focal
0      347.252959  427.819360   605.909188
1      621.834862  641.629131   735.825691
2      622.262074  903.746689  3501.073660
3      621.834862  641.629131   713.840496
4      621.834862  641.987211   709.472695

Optionally, you can specify which percentile values shall be computed.

>>> momepy.percentile(buildings.area, knn5, q=[10, 90]).head()
               10            90
focal
0      123.769329    683.514930
1      564.160901   1009.158671
2      564.160901  11216.093578
3      564.160901    929.400353
4      564.160901    903.746689