diff --git a/everviz/pages/deltaplot.py b/everviz/pages/deltaplot.py
index 83d3615..f8cf20d 100644
--- a/everviz/pages/deltaplot.py
+++ b/everviz/pages/deltaplot.py
@@ -22,7 +22,7 @@ def _get_objective_delta_values(api, best_batch):
 
 
 def _get_summary_delta_values(api, best_batch):
-    summary = api.summary_values()
+    summary = api.summary_values().to_pandas()
     if summary.empty:
         return summary
     data = (
diff --git a/everviz/pages/summary_values.py b/everviz/pages/summary_values.py
index c01469d..f68b3c2 100644
--- a/everviz/pages/summary_values.py
+++ b/everviz/pages/summary_values.py
@@ -25,7 +25,7 @@ def _set_up_data_sources(api, keys=None):
     everviz_path = os.path.join(everest_folder, "everviz")
 
     # Make a table which statistics over the realizations.
-    summary_values = api.summary_values(keys=keys)
+    summary_values = api.summary_values(keys=keys).to_pandas()
 
     if summary_values.empty:
         return None
diff --git a/pyproject.toml b/pyproject.toml
index 8690a33..0a21cc2 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -45,7 +45,8 @@ test = [
     "dash[testing]",
     "pillow",
     "pytest-mock",
-    "sphinx"
+    "sphinx",
+    "polars"
 ]
 style = [
     "pre-commit",
diff --git a/tests/integration/delta_plot/test_summary_delta_plot_integration.py b/tests/integration/delta_plot/test_summary_delta_plot_integration.py
index 2a74161..b97afa3 100644
--- a/tests/integration/delta_plot/test_summary_delta_plot_integration.py
+++ b/tests/integration/delta_plot/test_summary_delta_plot_integration.py
@@ -1,6 +1,6 @@
 from datetime import datetime
 
-import pandas
+import polars as pl
 from everviz.pages.deltaplot import _get_summary_delta_values
 from everviz.plugins.delta_plot.delta_plot import DeltaPlot
 
@@ -23,7 +23,7 @@
 
 def test_summary_delta_plot_callback(app, dash_duo, mocker, caplog):
     mock_api = mocker.Mock()
-    mock_api.summary_values.return_value = pandas.DataFrame(_SUMMARY)
+    mock_api.summary_values.return_value = pl.DataFrame(_SUMMARY)
 
     mocker.patch(
         "everviz.plugins.delta_plot.delta_plot.get_data",
@@ -60,7 +60,7 @@ def test_summary_delta_plot_callback(app, dash_duo, mocker, caplog):
 
 def test_summary_delta_plot_empty_callback(app, dash_duo, mocker, caplog):
     mock_api = mocker.Mock()
-    mock_api.summary_values.return_value = pandas.DataFrame(_EMPTY_SUMMARY)
+    mock_api.summary_values.return_value = pl.DataFrame(_EMPTY_SUMMARY)
 
     mocker.patch(
         "everviz.plugins.delta_plot.delta_plot.get_data",
diff --git a/tests/unit/test_delta_plot_data.py b/tests/unit/test_delta_plot_data.py
index 868ffb4..bc3fb5b 100644
--- a/tests/unit/test_delta_plot_data.py
+++ b/tests/unit/test_delta_plot_data.py
@@ -1,7 +1,7 @@
 import os
 from datetime import datetime
 
-import pandas
+import polars as pl
 from everviz.pages.deltaplot import (
     _get_objective_delta_values,
     _get_summary_delta_values,
@@ -59,7 +59,7 @@ def test_objective_values_data_frame(mocker):
 def test_summary_values_data_frame(mocker):
     """Test for the correct layout and size of the objective values data frame."""
     mock_api = mocker.Mock()
-    mock_api.summary_values.return_value = pandas.DataFrame(_SUMMARY)
+    mock_api.summary_values.return_value = pl.DataFrame(_SUMMARY)
 
     summary_delta_values = _get_summary_delta_values(mock_api, 2)
 
@@ -77,7 +77,7 @@ def test_set_up_sources(mocker, tmpdir):
     mock_api = mocker.Mock()
     mock_api.objective_values = _OBJECTIVES
     mock_api.single_objective_values = _SINGLE_OBJECTIVES
-    mock_api.summary_values.return_value = pandas.DataFrame(_SUMMARY)
+    mock_api.summary_values.return_value = pl.DataFrame(_SUMMARY)
     mock_api.output_folder = tmpdir
 
     os.mkdir(os.path.join(tmpdir, "everviz"))
@@ -96,7 +96,7 @@ def test_delta_plot_layout_with_empty_summary(mocker, tmpdir):
     mock_api = mocker.Mock()
     mock_api.objective_values = _OBJECTIVES
     mock_api.single_objective_values = _SINGLE_OBJECTIVES
-    mock_api.summary_values.return_value = pandas.DataFrame(
+    mock_api.summary_values.return_value = pl.DataFrame(
         {"realization": [], "simulation": [], "date": [], "batch": []}
     )
     mock_api.output_folder = tmpdir
diff --git a/tests/unit/test_summary_plot.py b/tests/unit/test_summary_plot.py
index b016082..0c2ecb6 100644
--- a/tests/unit/test_summary_plot.py
+++ b/tests/unit/test_summary_plot.py
@@ -2,6 +2,7 @@
 
 import numpy as np
 import pandas as pd
+import polars as pl
 from everviz.pages.summary_values import _summary_values, page_layout
 from everviz.plugins.summary_plot.util import calculate_statistics
 
@@ -32,7 +33,7 @@ def test_empty_summary_values(mocker):
     """Test for the case of missing summary values."""
     mock_api = mocker.Mock()
     mock_api.output_folder = "dummy"
-    mock_api.summary_values.return_value = pd.DataFrame()
+    mock_api.summary_values.return_value = pl.DataFrame()
     layout = page_layout(mock_api)
     assert not layout