Add option in raster plot to crop around centroids (#1047)

* add option to crop raster plots around centroids

* update docstrings

* change name of keyword parameter

* change keyword name and improve docstring descriptions

* updated changelog

* change default value of mask parameter and update changelog

Author: ValentinGebhart

CHANGELOG.md

@@ -15,7 +15,8 @@ Removed:
 - `pandas-datareader`
 
 ### Added
-
+- Added optional parameter to `geo_im_from_array`, `plot_from_gdf`, `plot_rp_imp`, `plot_rp_intensity`,
+`plot_intensity`, `plot_fraction`, `_event_plot` to mask plotting when regions are too far from data points [#1047](https://github.com/CLIMADA-project/climada_python/pull/1047). To recreate previous plots (no masking), the parameter can be set to None.
 - Added instructions to install Climada petals on Euler cluster in `doc.guide.Guide_Euler.ipynb` [#1029](https://github.com/CLIMADA-project/climada_python/pull/1029)
 
 ### Changed

climada/engine/impact.py

@@ -1178,6 +1178,7 @@ def plot_rp_imp(
         return_periods=(25, 50, 100, 250),
         log10_scale=True,
         axis=None,
+        mask_distance=0.01,
         kwargs_local_exceedance_impact=None,
         **kwargs,
     ):
@@ -1194,6 +1195,11 @@ def plot_rp_imp(
             plot impact as log10(impact). Default: True
         smooth : bool, optional
             smooth plot to plot.RESOLUTIONxplot.RESOLUTION. Default: True
+        mask_distance: float, optional
+            Only regions are plotted that are closer to any of the data points than this distance,
+            relative to overall plot size. For instance, to only plot values
+            at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+            Default is 0.01.
         kwargs_local_exceedance_impact: dict
             Dictionary of keyword arguments for the method impact.local_exceedance_impact.
         kwargs : dict, optional
@@ -1242,7 +1248,12 @@ def plot_rp_imp(
             )
 
         axis = u_plot.plot_from_gdf(
-            impacts_stats, title, column_labels, axis=axis, **kwargs
+            impacts_stats,
+            title,
+            column_labels,
+            axis=axis,
+            mask_distance=mask_distance,
+            **kwargs,
         )
         return axis, impacts_stats_vals
 

climada/hazard/plot.py

@@ -40,6 +40,7 @@ def plot_rp_intensity(
         self,
         return_periods=(25, 50, 100, 250),
         axis=None,
+        mask_distance=0.01,
         kwargs_local_exceedance_intensity=None,
         **kwargs,
     ):
@@ -56,6 +57,11 @@ def plot_rp_intensity(
             axis to use
         kwargs_local_exceedance_intensity: dict
             Dictionary of keyword arguments for the method hazard.local_exceedance_intensity.
+        mask_distance: float, optional
+            Only regions are plotted that are closer to any of the data points than this distance,
+            relative to overall plot size. For instance, to only plot values
+            at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+            Default is 0.01.
         kwargs: optional
             arguments for pcolormesh matplotlib function used in event plots
 
@@ -89,7 +95,12 @@ def plot_rp_intensity(
         )
 
         axis = u_plot.plot_from_gdf(
-            inten_stats, title, column_labels, axis=axis, **kwargs
+            inten_stats,
+            title,
+            column_labels,
+            axis=axis,
+            mask_distance=mask_distance,
+            **kwargs,
         )
         return axis, inten_stats.values[:, 1:].T.astype(float)
 
@@ -100,6 +111,7 @@ def plot_intensity(
         smooth=True,
         axis=None,
         adapt_fontsize=True,
+        mask_distance=0.01,
         **kwargs,
     ):
         """Plot intensity values for a selected event or centroid.
@@ -123,6 +135,11 @@ def plot_intensity(
             in module `climada.util.plot`)
         axis: matplotlib.axes._subplots.AxesSubplot, optional
             axis to use
+        mask_distance: float, optional
+            Only regions are plotted that are closer to any of the data points than this distance,
+            relative to overall plot size. For instance, to only plot values
+            at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+            Default is 0.01.
         kwargs: optional
             arguments for pcolormesh matplotlib function
             used in event plots or for plot function used in centroids plots
@@ -148,6 +165,7 @@ def plot_intensity(
                 crs_epsg,
                 axis,
                 adapt_fontsize=adapt_fontsize,
+                mask_distance=mask_distance,
                 **kwargs,
             )
         if centr is not None:
@@ -157,7 +175,15 @@ def plot_intensity(
 
         raise ValueError("Provide one event id or one centroid id.")
 
-    def plot_fraction(self, event=None, centr=None, smooth=True, axis=None, **kwargs):
+    def plot_fraction(
+        self,
+        event=None,
+        centr=None,
+        smooth=True,
+        axis=None,
+        mask_distance=0.01,
+        **kwargs,
+    ):
         """Plot fraction values for a selected event or centroid.
 
         Parameters
@@ -179,6 +205,11 @@ def plot_fraction(self, event=None, centr=None, smooth=True, axis=None, **kwargs
             in module `climada.util.plot`)
         axis: matplotlib.axes._subplots.AxesSubplot, optional
             axis to use
+        mask_distance: float, optional
+            Relative distance (with respect to maximal map extent in longitude or latitude) to data
+            points above which plot should not display values. For instance, to only plot values
+            at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+            Default is None.
         kwargs: optional
             arguments for pcolormesh matplotlib function
             used in event plots or for plot function used in centroids plots
@@ -196,7 +227,13 @@ def plot_fraction(self, event=None, centr=None, smooth=True, axis=None, **kwargs
             if isinstance(event, str):
                 event = self.get_event_id(event)
             return self._event_plot(
-                event, self.fraction, col_label, smooth, axis, **kwargs
+                event,
+                self.fraction,
+                col_label,
+                smooth,
+                axis,
+                mask_distance=mask_distance,
+                **kwargs,
             )
         if centr is not None:
             if isinstance(centr, tuple):
@@ -215,6 +252,7 @@ def _event_plot(
         axis=None,
         figsize=(9, 13),
         adapt_fontsize=True,
+        mask_distance=0.01,
         **kwargs,
     ):
         """Plot an event of the input matrix.
@@ -236,6 +274,11 @@ def _event_plot(
             axis to use
         figsize: tuple, optional
             figure size for plt.subplots
+        mask_distance: float, optional
+            Only regions are plotted that are closer to any of the data points than this distance,
+            relative to overall plot size. For instance, to only plot values
+            at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+            Default is None.
         kwargs: optional
             arguments for pcolormesh matplotlib function
 
@@ -283,6 +326,7 @@ def _event_plot(
             figsize=figsize,
             proj=crs_espg,
             adapt_fontsize=adapt_fontsize,
+            mask_distance=mask_distance,
             **kwargs,
         )
 

climada/util/plot.py

@@ -47,6 +47,7 @@
 from mpl_toolkits.axes_grid1 import make_axes_locatable
 from rasterio.crs import CRS
 from scipy.interpolate import griddata
+from scipy.spatial import cKDTree
 from shapely.geometry import box
 
 import climada.util.coordinates as u_coord
@@ -337,6 +338,7 @@ def geo_im_from_array(
     axes=None,
     figsize=(9, 13),
     adapt_fontsize=True,
+    mask_distance=0.01,
     **kwargs,
 ):
     """Image(s) plot defined in array(s) over input coordinates.
@@ -368,6 +370,11 @@ def geo_im_from_array(
     adapt_fontsize : bool, optional
         If set to true, the size of the fonts will be adapted to the size of the figure. Otherwise
         the default matplotlib font size is used. Default is True.
+    mask_distance: float, optional
+        Only regions are plotted that are closer to any of the data points than this distance,
+        relative to overall plot size. For instance, to only plot values
+        at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+        Default is 0.01.
     **kwargs
         arbitrary keyword arguments for pcolormesh matplotlib function
 
@@ -448,6 +455,17 @@ def geo_im_from_array(
                 (grid_x, grid_y),
                 fill_value=min_value,
             )
+            # Compute distance of each grid point to the nearest known point
+            if mask_distance is not None:
+                tree = cKDTree(np.array((coord[:, 1], coord[:, 0])).T)
+                distances, _ = tree.query(
+                    np.c_[grid_x.ravel(), grid_y.ravel()],
+                    p=2,  # for plotting squares and not sphere around centroids use p=np.inf
+                )
+                threshold = (
+                    max(extent[1] - extent[0], extent[3] - extent[2]) * mask_distance
+                )
+                grid_im[(distances.reshape(grid_im.shape) > threshold)] = min_value
         else:
             grid_x = coord[:, 1].reshape((width, height)).transpose()
             grid_y = coord[:, 0].reshape((width, height)).transpose()
@@ -477,7 +495,7 @@ def geo_im_from_array(
         )
         # handle NaNs in griddata
         color_nan = "gainsboro"
-        if np.any(np.isnan(x) for x in grid_im):
+        if np.isnan(grid_im).any():
             no_data_patch = mpatches.Patch(
                 facecolor=color_nan, edgecolor="black", label="NaN"
             )
@@ -1086,6 +1104,7 @@ def plot_from_gdf(
     axis=None,
     figsize=(9, 13),
     adapt_fontsize=True,
+    mask_distance=0.01,
     **kwargs,
 ):
     """Plot several subplots from different columns of a GeoDataFrame, e.g., for
@@ -1108,6 +1127,11 @@ def plot_from_gdf(
     adapt_fontsize: bool, optional
         If set to true, the size of the fonts will be adapted to the size of the figure.
         Otherwise the default matplotlib font size is used. Default is True.
+    mask_distance: float, optional
+        Relative distance (with respect to maximal map extent in longitude or latitude) to data
+        points above which plot should not display values. For instance, to only plot values
+        at the centroids, use mask_distance=0.01. If None, the plot is not masked.
+        Default is 0.01.
     kwargs: optional
         Arguments for pcolormesh matplotlib function used in event plots.
 
@@ -1168,6 +1192,7 @@ def plot_from_gdf(
         axes=axis,
         figsize=figsize,
         adapt_fontsize=adapt_fontsize,
+        mask_distance=mask_distance,
         **kwargs,
     )