diff --git a/_doc/examples/README.txt b/_doc/examples/README.txt
new file mode 100644
index 0000000..566f0e8
--- /dev/null
+++ b/_doc/examples/README.txt
@@ -0,0 +1,3 @@
+====================
+Galleries d'exemples
+====================
diff --git a/_doc/examples/lectures/README.txt b/_doc/examples/lectures/README.txt
new file mode 100644
index 0000000..962ff70
--- /dev/null
+++ b/_doc/examples/lectures/README.txt
@@ -0,0 +1,6 @@
+
+.. _l-gallery-examples:
+
+Exemples longs
+--------------
+
diff --git a/_doc/examples/lectures/plot_nuage.py b/_doc/examples/lectures/plot_nuage.py
new file mode 100644
index 0000000..6c86f96
--- /dev/null
+++ b/_doc/examples/lectures/plot_nuage.py
@@ -0,0 +1,42 @@
+# -*- coding: utf-8 -*-
+"""
+Corrélations
+============
+
+Dessine les corrélations pour un jeu de données.
+"""
+
+###############
+# A remplacer.
+
+import os
+import sys
+this = os.path.abspath(os.path.dirname(__file__))
+if "papierstat" in this:
+    this = this.split("papierstat")[0].rstrip("\\/")
+for module in ["papierstat"]:
+    try:
+        exec("import %s" % module)
+    except ImportError:
+        p = os.path.join(this, module, "src")
+        sys.path.append(p)
+        exec("import %s" % module)
+
+
+#########################
+# Récupération des données
+
+from papierstat.datasets import load_wines_dataset
+df = load_wines_dataset()
+print(df.head(n=2).T)
+
+####################
+# Les corrélations avec :epkg:`seaborn`.
+
+from seaborn import clustermap
+
+clustermap(df.corr(), center=0, cmap="vlag",
+           linewidths=.75, figsize=(13, 13))
+
+import matplotlib.pyplot as plt
+plt.show()
diff --git a/_doc/notebooks/lectures/wines_knn.ipynb b/_doc/notebooks/lectures/wines_knn.ipynb
index 4c3d1d7..36f5b18 100644
--- a/_doc/notebooks/lectures/wines_knn.ipynb
+++ b/_doc/notebooks/lectures/wines_knn.ipynb
@@ -175,8 +175,8 @@
         }
       ],
       "source": [
-        "from papierstat.datasets import load_wines_datasets\n",
-        "df = load_wines_datasets()\n",
+        "from papierstat.datasets import load_wines_dataset\n",
+        "df = load_wines_dataset()\n",
         "df.head()"
       ]
     },
diff --git a/_doc/notebooks/lectures/wines_knn_eval.ipynb b/_doc/notebooks/lectures/wines_knn_eval.ipynb
index 0ee6299..44710df 100644
--- a/_doc/notebooks/lectures/wines_knn_eval.ipynb
+++ b/_doc/notebooks/lectures/wines_knn_eval.ipynb
@@ -25,8 +25,8 @@
       "metadata": {},
       "outputs": [],
       "source": [
-        "from papierstat.datasets import load_wines_datasets\n",
-        "df = load_wines_datasets()\n",
+        "from papierstat.datasets import load_wines_dataset\n",
+        "df = load_wines_dataset()\n",
         "X = df.drop(['quality', 'color'], axis=1)\n",
         "y = df['quality']"
       ]
diff --git a/_doc/notebooks/lectures/wines_knn_split.ipynb b/_doc/notebooks/lectures/wines_knn_split.ipynb
index 950c0a0..d6701ed 100644
--- a/_doc/notebooks/lectures/wines_knn_split.ipynb
+++ b/_doc/notebooks/lectures/wines_knn_split.ipynb
@@ -24,8 +24,8 @@
       "metadata": {},
       "outputs": [],
       "source": [
-        "from papierstat.datasets import load_wines_datasets\n",
-        "df = load_wines_datasets()\n",
+        "from papierstat.datasets import load_wines_dataset\n",
+        "df = load_wines_dataset()\n",
         "X = df.drop(['quality', 'color'], axis=1)\n",
         "y = df['quality']"
       ]
diff --git a/_doc/sphinxdoc/source/biblio.rst b/_doc/sphinxdoc/source/biblio.rst
new file mode 100644
index 0000000..8a4e812
--- /dev/null
+++ b/_doc/sphinxdoc/source/biblio.rst
@@ -0,0 +1,15 @@
+
+Bibliographie
+=============
+
+*Anti-sèches*
+
+* `Anti-sèches de machine learning <http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx3/ii_antiseches.html>`_
+
+*Blog*
+
+* `Freakeconometrics <http://freakonometrics.hypotheses.org/>`_
+
+*Livres*
+
+* `The Elements of Statistical Learning <https://web.stanford.edu/~hastie/ElemStatLearn/>`_
diff --git a/_doc/sphinxdoc/source/index.rst b/_doc/sphinxdoc/source/index.rst
index 69fc278..4cd9389 100644
--- a/_doc/sphinxdoc/source/index.rst
+++ b/_doc/sphinxdoc/source/index.rst
@@ -18,6 +18,7 @@ Petit voyage au pays du machine learning
     introduction
     rappel
     lectures/index
+    biblio
     api/index
     all_notebooks
     i_ex
diff --git a/_doc/sphinxdoc/source/lectures/step1.rst b/_doc/sphinxdoc/source/lectures/step1.rst
index f2f09c5..e69c9ad 100644
--- a/_doc/sphinxdoc/source/lectures/step1.rst
+++ b/_doc/sphinxdoc/source/lectures/step1.rst
@@ -36,8 +36,8 @@ ci-dessous, pour deux vins.
     :rst:
 
     from pyquickhelper.pandashelper import df2rst
-    from papierstat.datasets import load_wines_datasets
-    df = load_wines_datasets()
+    from papierstat.datasets import load_wines_dataset
+    df = load_wines_dataset()
     print(df2rst(df.head(n=2).T.reset_index(drop=False)))
 
 On part du principe que si deux vins différents ont la même
@@ -75,14 +75,14 @@ Il peut être également obtenu avec le code suivant :
     :rst:
 
     from pyquickhelper.pandashelper import df2rst
-    from papierstat.datasets import load_wines_datasets
-    df = load_wines_datasets()
+    from papierstat.datasets import load_wines_dataset
+    df = load_wines_dataset()
     print(df2rst(df.head()))
 
 .. plot::
 
-    from papierstat.datasets import load_wines_datasets
-    df = load_wines_datasets()
+    from papierstat.datasets import load_wines_dataset
+    df = load_wines_dataset()
 
     import matplotlib.pyplot as plt
     plt.close('all')
diff --git a/_unittests/ut_datasets/test_wines.py b/_unittests/ut_datasets/test_wines.py
index 33313d2..ddb508c 100644
--- a/_unittests/ut_datasets/test_wines.py
+++ b/_unittests/ut_datasets/test_wines.py
@@ -39,7 +39,7 @@
 
 from pyquickhelper.loghelper import fLOG
 from pyquickhelper.pycode import ExtTestCase
-from src.papierstat.datasets import load_wines_datasets
+from src.papierstat.datasets import load_wines_dataset
 
 
 class TestWines(ExtTestCase):
@@ -50,7 +50,7 @@ def test_wines_download(self):
             self._testMethodName,
             OutputPrint=__name__ == "__main__")
 
-        df = load_wines_datasets(download=True)
+        df = load_wines_dataset(download=True)
         self.assertEqual(df.shape, (6497, 13))
 
     def test_wines_local(self):
@@ -59,7 +59,7 @@ def test_wines_local(self):
             self._testMethodName,
             OutputPrint=__name__ == "__main__")
 
-        df = load_wines_datasets(download=False)
+        df = load_wines_dataset(download=False)
         self.assertEqual(df.shape, (6497, 13))
 
 
diff --git a/requirements.txt b/requirements.txt
index 34a1e23..a2e4c9f 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -6,7 +6,7 @@ openpyxl
 pandas_streaming
 pycodestyle>=2.0.0
 pydocstyle
-pyquickhelper>=1.6.2283
+pyquickhelper>=1.6.2284
 sphinx-bootstrap-theme
 sphinx-gallery
 sphinx-redactor-theme
diff --git a/src/papierstat/datasets/__init__.py b/src/papierstat/datasets/__init__.py
index 0579412..b677644 100644
--- a/src/papierstat/datasets/__init__.py
+++ b/src/papierstat/datasets/__init__.py
@@ -16,7 +16,7 @@ def get_data_folder():
     return os.path.abspath(this)
 
 
-def load_wines_datasets(download=False):
+def load_wines_dataset(download=False):
     """
     Retourne le jeu de données
     `wines quality <https://archive.ics.uci.edu/ml/datasets/wine+quality>`_.