diff --git a/Orange/widgets/unsupervised/owlouvainclustering.py b/Orange/widgets/unsupervised/owlouvainclustering.py
index 3e6f1cecf9c..a572e1f3845 100644
--- a/Orange/widgets/unsupervised/owlouvainclustering.py
+++ b/Orange/widgets/unsupervised/owlouvainclustering.py
@@ -7,6 +7,7 @@
 from typing import Optional, Callable, Tuple, Any
 
 import numpy as np
+import scipy.sparse as sp
 import networkx as nx
 
 from AnyQt.QtCore import (
@@ -44,6 +45,16 @@
 METRICS = [("Euclidean", "l2"), ("Manhattan", "l1")]
 
 
+def is_same_data(t1: Table, t2: Table) -> bool:
+    """Check whether the `X`s in two tables are the same."""
+    # Sparse matrices can't be compared directly using np.array_equal otherwise
+    # a warning is triggered.
+    if t1.is_sparse() and t2.is_sparse():
+        return t1.X.shape == t2.X.shape and (t1.X != t2.X).nnz == 0
+    else:
+        return np.array_equal(t1.X, t2.X)
+
+
 class OWLouvainClustering(widget.OWWidget):
     name = "Louvain Clustering"
     description = "Detects communities in a network of nearest neighbors."
@@ -407,8 +418,7 @@ def set_data(self, data):
         # Make sure to properly enable/disable slider based on `apply_pca` setting
         self.controls.pca_components.setEnabled(self.apply_pca)
 
-        # If X hasn't changed, there's no reason to recompute clusters
-        if prev_data and self.data and np.array_equal(self.data.X, prev_data.X):
+        if prev_data and self.data and is_same_data(prev_data, self.data):
             if self.auto_commit:
                 self._send_data()
             return