diff --git a/superset/models/core.py b/superset/models/core.py
index c32ac5c229635..6eef48c688ccc 100644
--- a/superset/models/core.py
+++ b/superset/models/core.py
@@ -702,7 +702,7 @@ def needs_conversion(df_series):
                 return True
             return False
 
-        for k, v in df.dtypes.iteritems():
+        for k, v in df.dtypes.items():
             if v.type == numpy.object_ and needs_conversion(df[k]):
                 df[k] = df[k].apply(utils.json_dumps_w_dates)
         return df
diff --git a/superset/viz.py b/superset/viz.py
index 5a92cd3c59c17..cb20e1181ffbe 100644
--- a/superset/viz.py
+++ b/superset/viz.py
@@ -170,11 +170,21 @@ def get_df(self, query_obj=None):
                 if self.datasource.offset:
                     df[DTTM_ALIAS] += timedelta(hours=self.datasource.offset)
                 df[DTTM_ALIAS] += self.time_shift
+
+            self.df_metrics_to_num(df, query_obj.get('metrics') or [])
+
             df.replace([np.inf, -np.inf], np.nan)
             fillna = self.get_fillna_for_columns(df.columns)
             df = df.fillna(fillna)
         return df
 
+    @staticmethod
+    def df_metrics_to_num(df, metrics):
+        """Converting metrics to numeric when pandas.read_sql cannot"""
+        for col, dtype in df.dtypes.items():
+            if dtype.type == np.object_ and col in metrics:
+                df[col] = pd.to_numeric(df[col])
+
     def query_obj(self):
         """Building a query object"""
         form_data = self.form_data
@@ -1060,7 +1070,6 @@ def process_data(self, df, aggregate=False):
         df = df.fillna(0)
         if fd.get('granularity') == 'all':
             raise Exception(_('Pick a time granularity for your time series'))
-
         if not aggregate:
             df = df.pivot_table(
                 index=DTTM_ALIAS,
@@ -1384,7 +1393,7 @@ def get_data(self, df):
             pt = (pt / pt.sum()).T
         pt = pt.reindex(row.index)
         chart_data = []
-        for name, ys in pt.iteritems():
+        for name, ys in pt.items():
             if pt[name].dtype.kind not in 'biufc' or name in self.groupby:
                 continue
             if isinstance(name, string_types):
@@ -1395,7 +1404,7 @@ def get_data(self, df):
                 l = [str(s) for s in name[1:]]  # noqa: E741
                 series_title = ', '.join(l)
             values = []
-            for i, v in ys.iteritems():
+            for i, v in ys.items():
                 x = i
                 if isinstance(x, (tuple, list)):
                     x = ', '.join([text_type(s) for s in x])
diff --git a/tests/viz_tests.py b/tests/viz_tests.py
index 6822837e28312..a5adfc1f2540f 100644
--- a/tests/viz_tests.py
+++ b/tests/viz_tests.py
@@ -77,6 +77,8 @@ def test_get_df_handles_dttm_col(self):
         results.df.empty = False
         datasource.query = Mock(return_value=results)
         test_viz = viz.BaseViz(datasource, form_data)
+
+        test_viz.df_metrics_to_num = Mock()
         test_viz.get_fillna_for_columns = Mock(return_value=0)
         test_viz.get_df(query_obj)
         mock_call = df.__setitem__.mock_calls[0]