diff --git a/mapswipe_workers/tests/integration/test_overall_stats_creation.py b/mapswipe_workers/tests/integration/test_overall_stats_creation.py
new file mode 100644
index 00000000..6939d79f
--- /dev/null
+++ b/mapswipe_workers/tests/integration/test_overall_stats_creation.py
@@ -0,0 +1,130 @@
+
+import os
+import sys
+import types
+
+# Fake sentry_sdk
+fake_sentry = types.ModuleType("sentry_sdk")
+fake_sentry.init = lambda *args, **kwargs: None
+fake_sentry.capture_exception = lambda *args, **kwargs: None
+sys.modules["sentry_sdk"] = fake_sentry
+
+fake_xdg = types.ModuleType("xdg")
+fake_xdg.XDG_DATA_HOME = "/tmp"
+sys.modules["xdg"] = fake_xdg
+
+# Fake osgeo
+fake_osgeo = types.ModuleType("osgeo")
+fake_osgeo.ogr = types.ModuleType("ogr")
+fake_osgeo.osr = types.ModuleType("osr")
+sys.modules["osgeo"] = fake_osgeo
+
+# Fake env vars
+os.environ["FIREBASE_API_KEY"] = "dummy"
+os.environ["FIREBASE_AUTH_DOMAIN"] = "dummy"
+os.environ["FIREBASE_DATABASE_URL"] = "dummy"
+os.environ["FIREBASE_STORAGE_BUCKET"] = "dummy"
+os.environ["FIREBASE_PROJECT_ID"] = "dummy"
+os.environ["FIREBASE_MESSAGING_SENDER_ID"] = "dummy"
+os.environ["FIREBASE_APP_ID"] = "dummy"
+os.environ["POSTGRES_PASSWORD"] = "dummy"
+os.environ["OSMCHA_API_KEY"] = "dummy"
+os.environ["MAPILLARY_API_KEY"] = "dummy"
+
+
+import pandas as pd
+from tempfile import TemporaryDirectory
+
+from mapswipe_workers.generate_stats.overall_stats import get_overall_stats
+
+
+
+def test_overall_stats_csv_creation():
+    """
+    Just making sure that get_overall_stats() actually creates a CSV
+    and the numbers inside it make sense. Basically, testing that the
+    aggregation logic works the way it's supposed to.
+    """
+
+    # Fake project data to simulate what the real function would see.
+    df = pd.DataFrame(
+        [
+            {
+                "project_id": "p1",
+                "status": "active",
+                "area_sqkm": 10,
+                "number_of_results": 100,
+                "number_of_results_progress": 60,
+                "number_of_users": 5,
+            },
+            {
+                "project_id": "p2",
+                "status": "active",
+                "area_sqkm": 20,
+                "number_of_results": 200,
+                "number_of_results_progress": 140,
+                "number_of_users": 15,
+            },
+            {
+                "project_id": "p3",
+                "status": "finished",
+                "area_sqkm": 50,
+                "number_of_results": 300,
+                "number_of_results_progress": 250,
+                "number_of_users": 30,
+            },
+        ]
+    )
+
+    # Expected aggregated numbers for active projects.
+    expected_active_projects = 2
+    expected_active_area = 10 + 20
+    expected_active_results = 100 + 200
+    expected_active_results_progress = 60 + 140
+    expected_active_avg_users = (5 + 15) / 2
+
+    # Use a temp folder so we don't clutter anything in the repo.
+    with TemporaryDirectory() as tmpdir:
+        output_file = os.path.join(tmpdir, "overall_stats.csv")
+
+        # Run the actual function we are testing.
+        result_df = get_overall_stats(df, output_file)
+
+        # Check that the CSV file was actually created.
+        assert os.path.exists(output_file)
+
+        # Load the CSV that the function wrote out.
+        csv_df = pd.read_csv(output_file)
+
+        # These are the columns we expect to see.
+        expected_columns = [
+            "status",
+            "count_projects",
+            "area_sqkm",
+            "number_of_results",
+            "number_of_results_progress",
+            "average_number_of_users_per_project",
+        ]
+
+        # We sort here because pandas sometimes shuffles columns around,
+        assert sorted(csv_df.columns) == sorted(expected_columns)
+
+        # We expect exactly one row for "active" and one for "finished".
+        assert len(csv_df) == 2
+
+        # Grab the row for the active projects.
+        active_row = csv_df[csv_df["status"] == "active"].iloc[0]
+
+        # Now check actual math.
+        assert active_row["count_projects"] == expected_active_projects
+        assert active_row["area_sqkm"] == expected_active_area
+        assert active_row["number_of_results"] == expected_active_results
+        assert (
+            active_row["number_of_results_progress"]
+            == expected_active_results_progress
+        )
+        assert (
+            active_row["average_number_of_users_per_project"]
+            == expected_active_avg_users
+        )
+