Use pandas type checking for numeric dtype detection in Altair backend

Replace np.issubdtype() with pandas.api.types.is_numeric_dtype() to fix compatibility with pandas 3.0's new StringDtype.

NumPy's np.issubdtype() cannot handle pandas ExtensionDtype objects, particularly the new StringDtype introduced as default in pandas 3.0. This caused TypeError when checking if color columns were numeric: "Cannot interpret '<StringDtype(storage='python', na_value=nan)>' as a data type"

Using pandas' is_numeric_dtype() provides several benefits:
- Correctly handles all pandas dtypes (native NumPy and extension types)
- Works with numeric extension dtypes (Int64, Float64, Decimal, etc.)
- Compatible with both pandas 2.x and 3.x
- More semantically clear - checks if data is numeric rather than
  checking dtype inheritance hierarchy

Author: EwoutH

mesa/visualization/backends/altair_backend.py

@@ -276,7 +276,7 @@ def draw_agents(
         vmin = kwargs.pop("vmin", None)
         vmax = kwargs.pop("vmax", None)
 
-        color_is_numeric = np.issubdtype(df["original_color"].dtype, np.number)
+        color_is_numeric = pd.api.types.is_numeric_dtype(df["original_color"])
         if color_is_numeric:
             color_min = vmin if vmin is not None else df["original_color"].min()
             color_max = vmax if vmax is not None else df["original_color"].max()