diff --git a/.gitignore b/.gitignore
index 6e7f97b960..fe97af3e18 100644
--- a/.gitignore
+++ b/.gitignore
@@ -8,6 +8,7 @@
 # mpeltonen/sbt-idea plugin
 .idea_modules/
 .idea/
+.vscode/
 
 *.iml
 
diff --git a/databricks/koalas/extensions.py b/databricks/koalas/extensions.py
new file mode 100644
index 0000000000..ade7b29651
--- /dev/null
+++ b/databricks/koalas/extensions.py
@@ -0,0 +1,354 @@
+#
+# Copyright (C) 2019 Databricks, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+import warnings
+from functools import wraps
+
+
+class CachedAccessor:
+    """
+    Custom property-like object.
+
+    A descriptor for caching accessors:
+
+    Parameters
+    ----------
+    name : str
+        Namespace that accessor's methods, properties, etc will be accessed under, e.g. "foo" for a
+        dataframe accessor yields the accessor ``df.foo``
+    accessor: cls
+        Class with the extension methods.
+
+    Notes
+    -----
+    For accessor, the class's __init__ method assumes that you are registering an accessor for one
+    of ``Series``, ``DataFrame``, or ``Index``.
+
+    This object is not meant to be instantiated directly. Instead, use register_dataframe_accessor,
+    register_series_accessor, or register_index_accessor.
+
+    The Koalas accessor is modified based on pandas.core.accessor.
+    """
+
+    def __init__(self, name, accessor):
+        self._name = name
+        self._accessor = accessor
+
+    def __get__(self, obj, cls):
+        if obj is None:
+            return self._accessor
+        accessor_obj = self._accessor(obj)
+        setattr(obj, self._name, accessor_obj)
+        return accessor_obj
+
+
+def _register_accessor(name, cls):
+    """
+    Register a custom accessor on {klass} objects.
+
+    Parameters
+    ----------
+    name : str
+        Name under which the accessor should be registered. A warning is issued if this name
+        conflicts with a preexisting attribute.
+
+    Returns
+    -------
+    callable
+        A class decorator.
+
+    See Also
+    --------
+    register_dataframe_accessor: Register a custom accessor on DataFrame objects
+    register_series_accessor: Register a custom accessor on Series objects
+    register_index_accessor: Register a custom accessor on Index objects
+
+    Notes
+    -----
+    When accessed, your accessor will be initialiazed with the Koalas object the user is interacting
+    with. The code signature must be:
+
+    .. code-block:: python
+
+        def __init__(self, koalas_obj):
+            # constructor logic
+        ...
+
+    In the pandas API, if data passed to your accessor has an incorrect dtype, it's recommended to
+    raise an ``AttributeError`` for consistency purposes. In Koalas, ``ValueError`` is more
+    frequently used to annotate when a value's datatype is unexpected for a given method/function.
+
+    Ultimately, you can structure this however you like, but Koalas would likely do something like
+    this:
+
+    >>> ks.Series(['a', 'b']).dt
+    ...
+    Traceback (most recent call last):
+        ...
+    ValueError: Cannot call DatetimeMethods on type StringType
+
+    Note: This function is not meant to be used directly - instead, use register_dataframe_accessor,
+    register_series_accessor, or register_index_accessor.
+    """
+
+    def decorator(accessor):
+        if hasattr(cls, name):
+            msg = "registration of accessor {0} under name {1} for type {2} is overriding \
+                a preexisting attribute with the same name.".format(
+                accessor, name, cls
+            )
+
+            warnings.warn(
+                msg, UserWarning, stacklevel=2,
+            )
+        setattr(cls, name, CachedAccessor(name, accessor))
+        return accessor
+
+    return decorator
+
+
+def register_dataframe_accessor(name):
+    """
+    Register a custom accessor with a DataFrame
+
+    Parameters
+    ----------
+    name : str
+        name used when calling the accessor after its registered
+
+    Returns
+    -------
+    callable
+        A class decorator.
+
+    See Also
+    --------
+    register_series_accessor: Register a custom accessor on Series objects
+    register_index_accessor: Register a custom accessor on Index objects
+
+    Notes
+    -----
+    When accessed, your accessor will be initialiazed with the Koalas object the user is interacting
+    with. The accessor's init method should always ingest the object being accessed. See the
+    examples for the init signature.
+
+    In the pandas API, if data passed to your accessor has an incorrect dtype, it's recommended to
+    raise an ``AttributeError`` for consistency purposes. In Koalas, ``ValueError`` is more
+    frequently used to annotate when a value's datatype is unexpected for a given method/function.
+
+    Ultimately, you can structure this however you like, but Koalas would likely do something like
+    this:
+
+    >>> ks.Series(['a', 'b']).dt
+    ...
+    Traceback (most recent call last):
+        ...
+    ValueError: Cannot call DatetimeMethods on type StringType
+
+    Examples
+    --------
+    In your library code::
+
+        import databricks.koalas as ks
+
+        @ks.extensions.register_dataframe_accessor("geo")
+        class GeoAccessor:
+
+            def __init__(self, koalas_obj):
+                self._obj = koalas_obj
+                # other constructor logic
+
+            @property
+            def center(self):
+                # return the geographic center point of this DataFrame
+                lat = self._obj.latitude
+                lon = self._obj.longitude
+                return (float(lon.mean()), float(lat.mean()))
+
+            def plot(self):
+                # plot this array's data on a map
+                pass
+            ...
+
+    Then, in an ipython session::
+
+        >>> import databricks.koalas as ks
+        >>> from my_ext_lib import GeoAccessor # doctest: +SKIP
+        >>> type(GeoAccessor) # doctest: +SKIP
+            <class 'databricks.koalas.extensions.GeoAccessor'>
+        >>> kdf = ks.DataFrame({"longitude": np.linspace(0,10),
+        ...                     "latitude": np.linspace(0, 20)})
+        >>> kdf.geo.center # doctest: +SKIP
+            (5.0, 10.0)
+
+        >>> kdf.geo.plot() # doctest: +SKIP
+        ...
+    """
+    from databricks.koalas import DataFrame
+
+    return _register_accessor(name, DataFrame)
+
+
+def register_series_accessor(name):
+    """
+    Register a custom accessor with a Series object
+
+    Parameters
+    ----------
+    name : str
+        name used when calling the accessor after its registered
+
+    Returns
+    -------
+    callable
+        A class decorator.
+
+    See Also
+    --------
+    register_dataframe_accessor: Register a custom accessor on DataFrame objects
+    register_index_accessor: Register a custom accessor on Index objects
+
+    Notes
+    -----
+    When accessed, your accessor will be initialiazed with the Koalas object the user is interacting
+    with. The code signature must be::
+
+        def __init__(self, koalas_obj):
+            # constructor logic
+        ...
+
+    In the pandas API, if data passed to your accessor has an incorrect dtype, it's recommended to
+    raise an ``AttributeError`` for consistency purposes. In Koalas, ``ValueError`` is more
+    frequently used to annotate when a value's datatype is unexpected for a given method/function.
+
+    Ultimately, you can structure this however you like, but Koalas would likely do something like
+    this:
+
+    >>> ks.Series(['a', 'b']).dt
+    ...
+    Traceback (most recent call last):
+        ...
+    ValueError: Cannot call DatetimeMethods on type StringType
+
+    Examples
+    --------
+    In your library code::
+
+        import databricks.koalas as ks
+
+        @ks.extensions.register_series_accessor("geo")
+        class GeoAccessor:
+
+            def __init__(self, koalas_obj):
+                self._obj = koalas_obj
+
+            @property
+            def is_valid(self):
+                # boolean check to see if series contains valid geometry
+                return True if my_validation_logic(self._obj) else False
+            ...
+
+    Then, in an ipython session::
+
+        >>> import databricks.koalas as ks
+        >>> from my_ext_lib import GeoAccessor # doctest: +SKIP
+        >>> type(GeoAccessor) # doctest: +SKIP
+            <class 'databricks.koalas.extensions.GeoAccessor'>
+        >>> kdf = ks.DataFrame({"longitude": np.linspace(0,10),
+        ...                     "latitude": np.linspace(0, 20)})
+        >>> kdf.longitude.geo.is_valid # doctest: +SKIP
+            True
+        ...
+    """
+    from databricks.koalas import Series
+
+    return _register_accessor(name, Series)
+
+
+def register_index_accessor(name):
+    """
+    Register a custom accessor with an Index
+
+    Parameters
+    ----------
+    name : str
+        name used when calling the accessor after its registered
+
+    Returns
+    -------
+    callable
+        A class decorator.
+
+    See Also
+    --------
+    register_dataframe_accessor: Register a custom accessor on DataFrame objects
+    register_series_accessor: Register a custom accessor on Series objects
+
+    Notes
+    -----
+    When accessed, your accessor will be initialiazed with the Koalas object the user is interacting
+    with. The code signature must be::
+
+        def __init__(self, koalas_obj):
+            # constructor logic
+        ...
+
+    In the pandas API, if data passed to your accessor has an incorrect dtype, it's recommended to
+    raise an ``AttributeError`` for consistency purposes. In Koalas, ``ValueError`` is more
+    frequently used to annotate when a value's datatype is unexpected for a given method/function.
+
+    Ultimately, you can structure this however you like, but Koalas would likely do something like
+    this:
+
+    >>> ks.Series(['a', 'b']).dt
+    ...
+    Traceback (most recent call last):
+        ...
+    ValueError: Cannot call DatetimeMethods on type StringType
+
+    Examples
+    --------
+    In your library code::
+
+        import databricks.koalas as ks
+
+        @ks.extensions.register_series_accessor("foo")
+        class CustomAccessor:
+
+            def __init__(self, koalas_obj):
+                self._obj = koalas_obj
+                self.item = "baz"
+
+            @property
+            def bar(self):
+                # return item value
+                return self.item
+            ...
+
+    Then, in an ipython session::
+
+        >>> import databricks.koalas as ks
+        >>> from my_ext_lib import CustomAccessor # doctest: +SKIP
+        >>> type(CustomAccessor) # doctest: +SKIP
+            <class 'databricks.koalas.extensions.CustomAccessor'>
+        >>> kdf = ks.DataFrame({"longitude": np.linspace(0,10),
+        ...                     "latitude": np.linspace(0, 20)})
+        >>> kdf.index.foo.bar # doctest: +SKIP
+            "baz"
+        ...
+    """
+    from databricks.koalas import Index
+
+    return _register_accessor(name, Index)
diff --git a/databricks/koalas/testing/utils.py b/databricks/koalas/testing/utils.py
index 53bb345099..8c0c039cea 100644
--- a/databricks/koalas/testing/utils.py
+++ b/databricks/koalas/testing/utils.py
@@ -18,6 +18,7 @@
 import shutil
 import tempfile
 import unittest
+import warnings
 from contextlib import contextmanager
 
 import pandas as pd
@@ -282,3 +283,106 @@ def wrapped(self):
             compare(result_pandas, result_spark.toPandas())
 
     return wrapped
+
+
+@contextmanager
+def assert_produces_warning(
+    expected_warning=Warning,
+    filter_level="always",
+    check_stacklevel=True,
+    raise_on_extra_warnings=True,
+):
+    """
+    Context manager for running code expected to either raise a specific
+    warning, or not raise any warnings. Verifies that the code raises the
+    expected warning, and that it does not raise any other unexpected
+    warnings. It is basically a wrapper around ``warnings.catch_warnings``.
+
+    Notes
+    -----
+    Replicated from pandas._testing.
+
+    Parameters
+    ----------
+    expected_warning : {Warning, False, None}, default Warning
+        The type of Exception raised. ``exception.Warning`` is the base
+        class for all warnings. To check that no warning is returned,
+        specify ``False`` or ``None``.
+    filter_level : str or None, default "always"
+        Specifies whether warnings are ignored, displayed, or turned
+        into errors.
+        Valid values are:
+        * "error" - turns matching warnings into exceptions
+        * "ignore" - discard the warning
+        * "always" - always emit a warning
+        * "default" - print the warning the first time it is generated
+          from each location
+        * "module" - print the warning the first time it is generated
+          from each module
+        * "once" - print the warning the first time it is generated
+    check_stacklevel : bool, default True
+        If True, displays the line that called the function containing
+        the warning to show were the function is called. Otherwise, the
+        line that implements the function is displayed.
+    raise_on_extra_warnings : bool, default True
+        Whether extra warnings not of the type `expected_warning` should
+        cause the test to fail.
+
+    Examples
+    --------
+    >>> import warnings
+    >>> with assert_produces_warning():
+    ...     warnings.warn(UserWarning())
+    ...
+    >>> with assert_produces_warning(False): # doctest: +SKIP
+    ...     warnings.warn(RuntimeWarning())
+    ...
+    Traceback (most recent call last):
+        ...
+    AssertionError: Caused unexpected warning(s): ['RuntimeWarning'].
+    >>> with assert_produces_warning(UserWarning): # doctest: +SKIP
+    ...     warnings.warn(RuntimeWarning())
+    Traceback (most recent call last):
+        ...
+    AssertionError: Did not see expected warning of class 'UserWarning'
+    ..warn:: This is *not* thread-safe.
+    """
+    __tracebackhide__ = True
+
+    with warnings.catch_warnings(record=True) as w:
+
+        saw_warning = False
+        warnings.simplefilter(filter_level)
+        yield w
+        extra_warnings = []
+
+        for actual_warning in w:
+            if expected_warning and issubclass(actual_warning.category, expected_warning):
+                saw_warning = True
+
+                if check_stacklevel and issubclass(
+                    actual_warning.category, (FutureWarning, DeprecationWarning)
+                ):
+                    from inspect import getframeinfo, stack
+
+                    caller = getframeinfo(stack()[2][0])
+                    msg = (
+                        "Warning not set with correct stacklevel. ",
+                        "File where warning is raised: {} != ".format(actual_warning.filename),
+                        "{}. Warning message: {}".format(caller.filename, actual_warning.message),
+                    )
+                    assert actual_warning.filename == caller.filename, msg
+            else:
+                extra_warnings.append(
+                    (
+                        actual_warning.category.__name__,
+                        actual_warning.message,
+                        actual_warning.filename,
+                        actual_warning.lineno,
+                    )
+                )
+        if expected_warning:
+            msg = "Did not see expected warning of class {}".format(repr(expected_warning.__name__))
+            assert saw_warning, msg
+        if raise_on_extra_warnings and extra_warnings:
+            raise AssertionError("Caused unexpected warning(s): {}".format(repr(extra_warnings)))
diff --git a/databricks/koalas/tests/test_extension.py b/databricks/koalas/tests/test_extension.py
new file mode 100644
index 0000000000..e48562dae1
--- /dev/null
+++ b/databricks/koalas/tests/test_extension.py
@@ -0,0 +1,139 @@
+#
+# Copyright (C) 2019 Databricks, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+import contextlib
+import unittest
+import warnings
+
+import numpy as np
+import pandas as pd
+
+from databricks import koalas as ks
+from databricks.koalas.testing.utils import assert_produces_warning, ReusedSQLTestCase
+from databricks.koalas.extensions import (
+    register_dataframe_accessor,
+    register_series_accessor,
+    register_index_accessor,
+)
+
+
+@contextlib.contextmanager
+def ensure_removed(obj, attr):
+    """
+    Ensure attribute attached to 'obj' during testing is removed in the end
+    """
+    try:
+        yield
+
+    finally:
+        try:
+            delattr(obj, attr)
+        except AttributeError:
+            pass
+
+
+class CustomAccessor:
+    def __init__(self, obj):
+        self.obj = obj
+        self.item = "item"
+
+    @property
+    def prop(self):
+        return self.item
+
+    def method(self):
+        return self.item
+
+    def check_length(self, col=None):
+        if type(self.obj) == ks.DataFrame or col is not None:
+            return len(self.obj[col])
+        else:
+            try:
+                return len(self.obj)
+            except Exception as e:
+                raise ValueError(str(e))
+
+
+class ExtensionTest(ReusedSQLTestCase):
+    @property
+    def pdf(self):
+        return pd.DataFrame(
+            {"a": [1, 2, 3, 4, 5, 6, 7, 8, 9], "b": [4, 5, 6, 3, 2, 1, 0, 0, 0],},
+            index=np.random.rand(9),
+        )
+
+    @property
+    def kdf(self):
+        return ks.from_pandas(self.pdf)
+
+    @property
+    def accessor(self):
+        return CustomAccessor(self.kdf)
+
+    def test_setup(self):
+        self.assertEqual("item", self.accessor.item)
+
+    def test_dataframe_register(self):
+        with ensure_removed(ks.DataFrame, "test"):
+            register_dataframe_accessor("test")(CustomAccessor)
+            assert self.kdf.test.prop == "item"
+            assert self.kdf.test.method() == "item"
+            assert len(self.kdf["a"]) == self.kdf.test.check_length("a")
+
+    def test_series_register(self):
+        with ensure_removed(ks.Series, "test"):
+            register_series_accessor("test")(CustomAccessor)
+            assert self.kdf.a.test.prop == "item"
+            assert self.kdf.a.test.method() == "item"
+            assert self.kdf.a.test.check_length() == len(self.kdf["a"])
+
+    def test_index_register(self):
+        with ensure_removed(ks.Index, "test"):
+            register_index_accessor("test")(CustomAccessor)
+            assert self.kdf.index.test.prop == "item"
+            assert self.kdf.index.test.method() == "item"
+            assert self.kdf.index.test.check_length() == self.kdf.index.size
+
+    def test_accessor_works(self):
+        register_series_accessor("test")(CustomAccessor)
+
+        s = ks.Series([1, 2])
+        assert s.test.obj is s
+        assert s.test.prop == "item"
+        assert s.test.method() == "item"
+
+    def test_overwrite_warns(self):
+        mean = ks.Series.mean
+        try:
+            with assert_produces_warning(UserWarning) as w:
+                register_series_accessor("mean")(CustomAccessor)
+                s = ks.Series([1, 2])
+                assert s.mean.prop == "item"
+            msg = str(w[0].message)
+            assert "mean" in msg
+            assert "CustomAccessor" in msg
+            assert "Series" in msg
+        finally:
+            ks.Series.mean = mean
+
+    def test_raises_attr_error(self):
+        with ensure_removed(ks.Series, "bad"):
+
+            class Bad:
+                def __init__(self, data):
+                    raise AttributeError("whoops")
+
+            with self.assertRaises(AttributeError):
+                ks.Series([1, 2], dtype=object).bad
diff --git a/docs/source/reference/extensions.rst b/docs/source/reference/extensions.rst
new file mode 100644
index 0000000000..1dae2b0da2
--- /dev/null
+++ b/docs/source/reference/extensions.rst
@@ -0,0 +1,21 @@
+.. _api.extensions:
+
+==========
+Extensions
+==========
+.. currentmodule:: databricks.koalas.extensions
+
+Accessors
+---------
+
+Accessors can be written and registered with Koalas Dataframes, Series, and
+Index objects. Accessors allow developers to extend the functionality of
+Koalas objects seamlessly by writing arbitrary classes and methods which are
+then wrapped in one of the following decorators.
+
+.. autosummary::
+   :toctree: api/
+
+   register_dataframe_accessor
+   register_series_accessor
+   register_index_accessor
\ No newline at end of file
diff --git a/docs/source/reference/index.rst b/docs/source/reference/index.rst
index 34a6f00313..7258a3dd4a 100644
--- a/docs/source/reference/index.rst
+++ b/docs/source/reference/index.rst
@@ -12,3 +12,4 @@ API Reference
     window
     groupby
     ml
+    extensions