diff --git a/docs/changes.md b/docs/changes.md
index bea33d95..e8e02864 100644
--- a/docs/changes.md
+++ b/docs/changes.md
@@ -14,6 +14,7 @@
 - add `transformations.__call__()` to apply it
 - fix small bug in `pointcloud.distance_to()`
 - add `applications.MorphPlotter()` to morph a polygonal mesh to a target mesh
+- add `smooth_data()` to smooth/diffuse data attributes
 
 
 ## Breaking changes
@@ -40,6 +41,7 @@
 ## New/Revised Examples
 ```
 examples/advanced/warp4c.py
+examples/advanced/diffuse_data.py
 examples/volumetric/slab_vol.py
 examples/other/madcad1.py
 examples/other/tetgen1.py
diff --git a/docs/examples_db.js b/docs/examples_db.js
index 2431e5a8..8ee91934 100755
--- a/docs/examples_db.js
+++ b/docs/examples_db.js
@@ -702,6 +702,14 @@ vedo_example_db =
         long  : 'Interpolate cell values from a quad-mesh to a tri-mesh of different resolution',
         imgsrc: 'images/advanced/interpolateScalar4.png',
     },
+    {
+        pyname: 'diffuse_data',
+        kbd   : '',
+        categ : 'advanced',
+        short : 'smooth array',
+        long  : 'Smooth/diffuse an array of scalars on a mesh',
+        imgsrc: 'images/advanced/diffuse_data.png',
+    },
     {
         pyname: 'cut_with_mesh1',
         kbd   : '',
diff --git a/examples/advanced/diffuse_data.py b/examples/advanced/diffuse_data.py
new file mode 100644
index 00000000..28c7f057
--- /dev/null
+++ b/examples/advanced/diffuse_data.py
@@ -0,0 +1,36 @@
+import numpy as np
+from vedo import Grid, settings, show
+from vedo.pyplot import histogram
+
+settings.default_font = "FiraMonoMedium"
+
+grid = Grid(res=[50,50])
+grid.wireframe(False).lw(0)
+
+values = np.zeros(grid.npoints)
+values[int(grid.npoints/2)] = 1
+values[int(grid.npoints/5)] = 1
+
+grid.pointdata["scalars"] = values
+grid.cmap("Set1_r").add_scalarbar()
+
+grid2 = grid.clone()
+grid2.smooth_data(niter=750, relaxation_factor=0.1, strategy=1)
+grid2.cmap("Set1_r").add_scalarbar()
+
+his = histogram(
+    grid2.pointdata["scalars"],
+    c='k4',
+    xtitle="Concentration",
+    ytitle="Frequency",
+    axes=dict(htitle="", axes_linewidth=2, xyframe_line=0),
+)
+his = his.clone2d() # fix it to screen coords
+
+print("integrated over domain:", grid2.integrate_arrays_over_domain())
+
+show([
+    ["Initial state", grid],
+    ["After diffusion", grid2, his]],
+    N=2, axes=1,
+).close()
diff --git a/vedo/core.py b/vedo/core.py
index 9e4e7c5b..5b5beae4 100644
--- a/vedo/core.py
+++ b/vedo/core.py
@@ -1384,6 +1384,63 @@ def shrink(self, fraction=0.8):
         )
         return self
 
+    def smooth_data(self, 
+            niter=10, relaxation_factor=0.1, strategy=0, mask=None,
+            exclude=("Normals", "TextureCoordinates"),
+        ):
+        """
+        Smooth point attribute data using distance weighted Laplacian kernel.
+
+        The effect is to blur regions of high variation and emphasize low variation regions.
+
+        Arguments:
+            niter : (int)
+                number of iterations
+            relaxation_factor : (float)
+                relaxation factor controlling the amount of Laplacian smoothing applied
+            strategy : (int)
+                strategy to use for Laplacian smoothing
+                    - 0: use all points, all point data attributes are smoothed
+                    - 1: smooth all point attribute data except those on the boundary
+                    - 2: only point data connected to a boundary point are smoothed
+            mask : (str, np.ndarray)
+                array to be used as a mask (ignore then the strategy keyword)
+            exclude : (list)
+                list of arrays to be excluded from smoothing
+        """
+        saf = vtk.new("AttributeSmoothingFilter")
+        saf.SetInputData(self.dataset)
+        saf.SetRelaxationFactor(relaxation_factor)
+        saf.SetNumberOfIterations(niter)
+
+        for ex in exclude:
+            saf.AddExcludedArray(ex)
+
+        saf.SetWeightsTypeToDistance2()
+
+        saf.SetSmoothingStrategy(strategy)
+        if mask is not None:
+            saf.SetSmoothingStrategyToSmoothingMask()
+            if isinstance(mask, str):
+                mask_ = self.dataset.GetPointData().GetArray(mask)
+                if not mask_:
+                    vedo.logger.error(f"smooth_data(): mask array {mask} not found")
+                    return self
+                mask_array = vtk.vtkUnsignedCharArray()
+                mask_array.ShallowCopy(mask_)
+                mask_array.SetName(mask_.GetName())
+            else:
+                mask_array = utils.numpy2vtk(mask, dtype=np.uint8)
+            saf.SetSmoothingMask(mask_array)
+
+        saf.Update()
+
+        self._update(saf.GetOutput())
+        self.pipeline = utils.OperationNode(
+            "smooth_data", comment=f"strategy {strategy}", parents=[self], c="#9e2a2b"
+        )
+        return self
+        
 
 ###############################################################################
 class PointAlgorithms(CommonAlgorithms):
diff --git a/vedo/vtkclasses.py b/vedo/vtkclasses.py
index 4d7f8483..995d5d21 100644
--- a/vedo/vtkclasses.py
+++ b/vedo/vtkclasses.py
@@ -336,8 +336,9 @@
     location["vtkCellTreeLocator"] = "vtkFiltersGeneral"
 
 
-location["vtkGeometryFilter"] = "vtkFiltersGeometry"
+location["vtkAttributeSmoothingFilter"] = "vtkFiltersGeometry"
 location["vtkDataSetSurfaceFilter"] = "vtkFiltersGeometry"
+location["vtkGeometryFilter"] = "vtkFiltersGeometry"
 location["vtkImageDataGeometryFilter"] = "vtkFiltersGeometry"