diff --git a/TM1py/Services/CellService.py b/TM1py/Services/CellService.py
index 4210ba95..902f45f6 100644
--- a/TM1py/Services/CellService.py
+++ b/TM1py/Services/CellService.py
@@ -2798,13 +2798,15 @@ def extract_cellset_dataframe(
     @tidy_cellset
     @require_pandas
     def extract_cellset_dataframe_shaped(self, cellset_id: str, sandbox_name: str = None,
-                                         display_attribute: bool = False, **kwargs) -> 'pd.DataFrame':
+                                         display_attribute: bool = False, infer_dtype: bool = False, **kwargs) -> 'pd.DataFrame':
         """ Retrieves data from cellset in the shape of the query.
         Dimensions on rows can be stacked. One dimension must be placed on columns. Title selections are ignored.
 
         :param cellset_id
         :param sandbox_name: str
         :param display_attribute: bool, show element name or first attribute from MDX PROPERTIES clause
+        :param infer_dtype: bool, if True, lets pandas infer dtypes, otherwise all columns will be of type str.
+
         """
         url = "/api/v1/Cellsets('{}')?$expand=" \
               "Axes($filter=Ordinal eq 0 or Ordinal eq 1;$expand=Tuples(" \
@@ -2860,7 +2862,10 @@ def extract_cellset_dataframe_shaped(self, cellset_id: str, sandbox_name: str =
 
         for element_tuple, cells in zip(element_names_by_row, cell_values_by_row):
             body.append(list(element_tuple) + cells)
-        return pd.DataFrame(body, columns=headers, dtype=str)
+        if infer_dtype:
+            return pd.DataFrame(body, columns=headers)
+        else:
+            return pd.DataFrame(body, columns=headers, dtype=str)
 
     @require_pandas
     def extract_cellset_dataframe_pivot(self, cellset_id: str, dropna: bool = False, fill_value: bool = False,