diff --git a/docs/conf.py b/docs/conf.py
index bacd2eb..30b10df 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -57,6 +57,8 @@
     "sphinx.ext.todo",
     "sphinx_rtd_theme",
     "nbsphinx",
+    "sphinx.ext.graphviz",
+    "sphinx.ext.intersphinx",
 ]
 
 # Add any paths that contain templates here, relative to this directory.
@@ -67,6 +69,10 @@
 # This pattern also affects html_static_path and html_extra_path.
 exclude_patterns = ["_build", "Thumbs.db", ".DS_Store", "'**.ipynb_checkpoints'"]
 
+intersphinx_mapping = {
+    "python": ("https://python.readthedocs.io/en/latest", None),
+}
+
 
 # https://www.notion.so/Deepnote-Launch-Buttons-63c642a5e875463495ed2341e83a4b2a
 
diff --git a/docs/index.rst b/docs/index.rst
index 57cf9e3..8c093f5 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -42,6 +42,13 @@ datasets <building-introduction>`.
 **Using training datasets**
 
 -  :doc:`using/introduction`
+-  :doc:`using/opening`
+-  :doc:`using/subsetting`
+-  :doc:`using/combining`
+-  :doc:`using/selecting`
+-  :doc:`using/grids`
+-  :doc:`using/statistics`
+-  :doc:`using/other`
 -  :doc:`using/options`
 
 .. toctree::
@@ -50,6 +57,13 @@ datasets <building-introduction>`.
    :caption: Using datasets
 
    using/introduction
+   using/opening
+   using/subsetting
+   using/combining
+   using/selecting
+   using/grids
+   using/statistics
+   using/other
    using/options
 
 **Building training datasets**
diff --git a/docs/using/combining.rst b/docs/using/combining.rst
new file mode 100644
index 0000000..ed3d0ed
--- /dev/null
+++ b/docs/using/combining.rst
@@ -0,0 +1,81 @@
+.. _combining-datasets:
+
+####################
+ Combining datasets
+####################
+
+When combining datasets, the statistics of the first dataset are used by
+default. You can change this by setting the :ref:`selecting-statistics`
+option to a different dataset, even if it is not part of the
+combination. See
+
+.. _concat:
+
+********
+ concat
+********
+
+You can concatenate two or more datasets along the dates dimension. The
+package will check that all datasets are compatible (same resolution,
+same variables, etc.). Currently, the datasets must be given in
+chronological order with no gaps between them.
+
+.. code:: python
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1940-1978-1h-v2",
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+   )
+
+.. image:: concat.png
+   :alt: Concatenation
+
+Please note that you can pass more than two ``zarr`` files to the
+function.
+
+   **NOTE:** When concatenating file, the statistics are not recomputed;
+   it is the statistics of first file that are returned to the user.
+
+******
+ join
+******
+
+You can join two datasets that have the same dates, combining their
+variables.
+
+.. code:: python
+
+   from ecml_tools.data import open_dataset
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       "some-extra-parameters-from-another-source-o96-1979-2022-1h-v2",
+   )
+
+.. image:: join.png
+   :alt: Join
+
+If a variable is present in more that one file, that last occurrence of
+that variable will be used, and will be at the position of the first
+occurrence of that name.
+
+.. image:: overlay.png
+   :alt: Overlay
+
+Please note that you can join more than two ``zarr`` files.
+
+***********
+ ensembles
+***********
+
+.. code:: python
+
+   open_dataset(ensembles=[dataset1, dataset2, ...])
+
+*******
+ grids
+*******
+
+.. code:: python
+
+   open_dataset(grids=[dataset1, dataset2, ...], method=...)
diff --git a/docs/using/grids.rst b/docs/using/grids.rst
new file mode 100644
index 0000000..534320f
--- /dev/null
+++ b/docs/using/grids.rst
@@ -0,0 +1,17 @@
+.. _selecting-grids:
+
+#######################
+ Selecting grid points
+#######################
+
+**********
+ thinning
+**********
+
+.. code:: python
+
+   open_dataset(dataset, thinning=..., method="every-nth")
+
+******
+ area
+******
diff --git a/docs/using/opening.rst b/docs/using/opening.rst
new file mode 100644
index 0000000..9934728
--- /dev/null
+++ b/docs/using/opening.rst
@@ -0,0 +1,52 @@
+.. _opening-datasets:
+
+##################
+ Opening datasets
+##################
+
+.. code:: python
+
+   from anemoi_datasets import open_dataset
+
+   ds = open_dataset("path/to/dataset.zarr", option1=value1, option2=value2, ...)
+
+or
+
+.. code:: python
+
+   from anemoi_datasets import open_dataset
+
+   ds = open_dataset(combine=["path/to/dataset1.zarr",
+                              "path/to/dataset2.zarr", ...])
+
+or
+
+.. code:: python
+
+   from anemoi_datasets import open_dataset
+
+   ds = open_dataset(combine=["path/to/dataset1.zarr",
+                              "path/to/dataset2.zarr", ...],
+                              option1=value1, option2=value2, ...)
+
+The term `combine` is one of `join`, `concat`, `ensembles`, etc. See
+:ref:`combining-datasets` for more information.
+
+.. note::
+
+   The options `option1`, `option2`, apply to the combined dataset.
+
+.. code:: python
+
+   from anemoi_datasets import open_dataset
+
+   ds = open_dataset(combine=[{"dataset": "path/to/dataset1.zarr",
+                               "option1"=value1, "option2"=value2, ...},
+                              {"dataset": "path/to/dataset2.zarr",
+                               "option3"=value3, "option4"=value4, ...},
+                              ...])
+
+.. note::
+
+   The options `option1`, `option2`, apply to the first dataset, and
+   `option3`, `option4`, to the second dataset, etc.
diff --git a/docs/using/options.rst b/docs/using/options.rst
index 7659de8..e5953ea 100644
--- a/docs/using/options.rst
+++ b/docs/using/options.rst
@@ -71,7 +71,8 @@ dataset
 =======
 
 This is a path or URL to a ``zarr`` file that has been created with this
-package, as described in :ref:`datasets-building`.
+package, as described in :ref:`Building training datasets
+<building-introduction>`.
 
 .. code:: python
 
@@ -94,199 +95,3 @@ Alternatively, you can pass an already opened dataset:
 
    ds1 = open_dataset("aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2")
    ds2 = open_dataset(ds1, start=1979, end=2020)
-
-start
-=====
-
-This option let you subset the dataset by time. You can pass a date or a
-
-.. code:: python
-
-   open_dataset(dataset, start=1980)
-
-end
-===
-
-As for the start option, you can pass a date or a string:
-
-.. code:: python
-
-   open_dataset(dataset, end="2020-12-31")
-
-The following are equivalent way of describing ``start`` or ``end``:
-
--  ``2020`` and ``"2020"``
--  ``202306``, ``"202306"`` and ``"2023-06"``
--  ``20200301``, ``"20200301"`` and ``"2020-03-01"``
-
-frequency
-=========
-
-You can change the frequency of the dataset by passing a string with the
-
-.. code:: python
-
-   ds = open_dataset("aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2", frequency="6h")
-
-select
-======
-
-.. code:: python
-
-   # Select '2t' and 'tp' in that order
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       select=["2t", "tp"],
-   )
-
-.. code:: python
-
-   # Select '2t' and 'tp', but preserve the order in which they are in the dataset
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       select={"2t", "tp"},
-   )
-
-drop
-====
-
-You can also drop some variables:
-
-.. code:: python
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       drop=["10u", "10v"],
-   )
-
-reorder
-=======
-
-and reorder them:
-
-... using a list
-
-.. code:: python
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       reorder=["2t", "msl", "sp", "10u", "10v"],
-   )
-
-... or using a dictionary
-
-.. code:: python
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       reorder={"2t": 0, "msl": 1, "sp": 2, "10u": 3, "10v": 4},
-   )
-
-rename
-======
-
-You can also rename variables:
-
-.. code:: python
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       rename={"2t": "t2m"},
-   )
-
-This will be useful when your join datasets and do not want variables
-from one dataset to override the ones from the other.
-
-.. _statistics:
-
-statistics
-==========
-
-.. code:: python
-
-   open_dataset(dataset, statistics=other_dataset)
-
-thinning
-========
-
-.. code:: python
-
-   open_dataset(dataset, thinning=..., method="every-nth")
-
-area
-====
-
-********************
- Combining datasets
-********************
-
-When combining datasets, the statistics of the first dataset are used by
-default. You can change this by setting the :ref:`statistics` option to
-a different dataset, even if it is not part of the combination. See
-
-concat
-======
-
-You can concatenate two or more datasets along the dates dimension. The
-package will check that all datasets are compatible (same resolution,
-same variables, etc.). Currently, the datasets must be given in
-chronological order with no gaps between them.
-
-.. code:: python
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1940-1978-1h-v2",
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-   )
-
-.. image:: concat.png
-   :alt: Concatenation
-
-Please note that you can pass more than two ``zarr`` files to the
-function.
-
-   **NOTE:** When concatenating file, the statistics are not recomputed;
-   it is the statistics of first file that are returned to the user.
-
-join
-====
-
-You can join two datasets that have the same dates, combining their
-variables.
-
-.. code:: python
-
-   from ecml_tools.data import open_dataset
-
-   ds = open_dataset(
-       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
-       "some-extra-parameters-from-another-source-o96-1979-2022-1h-v2",
-   )
-
-.. image:: join.png
-   :alt: Join
-
-If a variable is present in more that one file, that last occurrence of
-that variable will be used, and will be at the position of the first
-occurrence of that name.
-
-.. image:: overlay.png
-   :alt: Overlay
-
-Please note that you can join more than two ``zarr`` files.
-
-ensembles
-=========
-
-.. code:: python
-
-   open_dataset(ensembles=[dataset1, dataset2, ...])
-
-grids
-=====
-
-.. code:: python
-
-   open_dataset(grids=[dataset1, dataset2, ...], method=...)
diff --git a/docs/using/other.rst b/docs/using/other.rst
new file mode 100644
index 0000000..3f45607
--- /dev/null
+++ b/docs/using/other.rst
@@ -0,0 +1,53 @@
+.. _selecting-other:
+
+##################
+ Other operations
+##################
+
+.. warning:: The operations described in this section are do not check that their inputs are compatible.
+
+
+*****
+ zip
+*****
+
+The `zip` operation is used to combine multiple datasets into a single dataset.
+
+.. code:: python
+
+   ds = open_dataset(zip=[dataset1, dataset2, ...])
+
+   # This will return tuples
+
+   print(ds[0])
+
+   print(ds[3, 4])
+
+
+
+This operation is identical to the Python's :py:func:`zip` function.
+
+*******
+ chain
+*******
+
+.. code:: python
+
+   ds = open_dataset(chain=[dataset1, dataset2, ...])
+
+
+The `chain` operation is used to combine multiple datasets into a single dataset.
+The datasets are combined by concatenating the data arrays along the first dimension (dates).
+This is similar to the :ref:`concat` operation, but no check are done to see if the datasets are compatible,
+this means that the shape of the arrays returned when iterating or indexing may be different.
+
+This operation is identical to the Python's :py:func:`itertools.chain` function.
+
+
+********
+ shuffle
+********
+
+.. code:: python
+
+   ds = open_dataset(dataset, shuffle=True)
diff --git a/docs/using/selecting.rst b/docs/using/selecting.rst
new file mode 100644
index 0000000..509b45d
--- /dev/null
+++ b/docs/using/selecting.rst
@@ -0,0 +1,80 @@
+.. _selecting-variables:
+
+#####################
+ Selecting variables
+#####################
+
+********
+ select
+********
+
+.. code:: python
+
+   # Select '2t' and 'tp' in that order
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       select=["2t", "tp"],
+   )
+
+.. code:: python
+
+   # Select '2t' and 'tp', but preserve the order in which they are in the dataset
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       select={"2t", "tp"},
+   )
+
+******
+ drop
+******
+
+You can also drop some variables:
+
+.. code:: python
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       drop=["10u", "10v"],
+   )
+
+*********
+ reorder
+*********
+
+and reorder them:
+
+... using a list
+
+.. code:: python
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       reorder=["2t", "msl", "sp", "10u", "10v"],
+   )
+
+... or using a dictionary
+
+.. code:: python
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       reorder={"2t": 0, "msl": 1, "sp": 2, "10u": 3, "10v": 4},
+   )
+
+********
+ rename
+********
+
+You can also rename variables:
+
+.. code:: python
+
+   ds = open_dataset(
+       "aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2",
+       rename={"2t": "t2m"},
+   )
+
+This will be useful when your join datasets and do not want variables
+from one dataset to override the ones from the other.
diff --git a/docs/using/statistics.rst b/docs/using/statistics.rst
new file mode 100644
index 0000000..e59f117
--- /dev/null
+++ b/docs/using/statistics.rst
@@ -0,0 +1,13 @@
+.. _selecting-statistics:
+
+############
+ Statistics
+############
+
+************
+ statistics
+************
+
+.. code:: python
+
+   open_dataset(dataset, statistics=other_dataset)
diff --git a/docs/using/subsetting.rst b/docs/using/subsetting.rst
new file mode 100644
index 0000000..137ab24
--- /dev/null
+++ b/docs/using/subsetting.rst
@@ -0,0 +1,41 @@
+.. _subsetting-datasets:
+
+#####################
+ Subsetting datasets
+#####################
+
+*******
+ start
+*******
+
+This option let you subset the dataset by time. You can pass a date or a
+
+.. code:: python
+
+   open_dataset(dataset, start=1980)
+
+*****
+ end
+*****
+
+As for the start option, you can pass a date or a string:
+
+.. code:: python
+
+   open_dataset(dataset, end="2020-12-31")
+
+The following are equivalent way of describing ``start`` or ``end``:
+
+-  ``2020`` and ``"2020"``
+-  ``202306``, ``"202306"`` and ``"2023-06"``
+-  ``20200301``, ``"20200301"`` and ``"2020-03-01"``
+
+***********
+ frequency
+***********
+
+You can change the frequency of the dataset by passing a string with the
+
+.. code:: python
+
+   ds = open_dataset("aifs-ea-an-oper-0001-mars-o96-1979-2022-1h-v2", frequency="6h")