diff --git a/docs/canvas/basics.md b/docs/canvas/basics.md
new file mode 100644
index 00000000..45892a2f
--- /dev/null
+++ b/docs/canvas/basics.md
@@ -0,0 +1,106 @@
+# Basics
+
+In `whitecanvas`, a "canvas" is an object with a layer list, axes elements, labels and
+methods to handle them. A "grid" is a collection of canvases and is implemented with
+rendering backends.
+
+## Create Canvases
+
+Canvas is created by [`new_canvas`][whitecanvas.core.new_canvas] method.
+
+- `backend`: the name of the backend to use.
+- `size`: the size of the canvas (width × height in pixels).
+- `palette`: the name of the color palette to use. Any input that is accepted by the
+  [cmap.Colormap](https://cmap-docs.readthedocs.io/en/latest/catalog/) can be used,
+  which includes:
+  1. The name of the built-in colormaps, such as "viridis", "plasma", etc.
+  2. Sequence of colors, such as `["red", "green", "blue"]`.
+  3. A `cmap.Colormap` object.
+
+``` python
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib", size=(400, 300), palette="tab10")
+```
+
+A `SingleCanvas` object is returned. It can be considered as a mixture of a canvas
+and a grid. All of the functions mentioned above are implemented in this object.
+Therefore, you can show the canvas by calling the `show` method.
+
+``` python
+#!skip
+canvas.show()  # show the canvas, depending on the backend
+```
+
+## Namespaces
+
+To make the API tidy, each element of a canvas is organized in namespaces. By using
+namespaces, we can avoid the long list of arguments and make the code more readable.
+For example, the x-limit of the canvas is controlled by the property `canvas.x.lim`.
+
+Following list shows the namespaces and the properties associated with them.
+
+- `canvas.{x, y}`: the x/y-axis.
+    - (*property*) `canvas.{x, y}.lim`: the x/y-limit as a tuple of floats.
+    - (*property*) `canvas.{x, y}.color`: the color of the x/y-axis.
+    - (*property*) `canvas.{x, y}.flipped`: whether the x/y-axis is flipped.
+    - `canvas.{x, y}.label`: the x/y-axis label.
+        - (*property*) `canvas.{x, y}.label.text`: the text of the label.
+        - (*property*) `canvas.{x, y}.label.color`: the color of the label as `ndarray`.
+        - (*property*) `canvas.{x, y}.label.size`: the font size of the label.
+        - (*property*) `canvas.{x, y}.label.family`: the font family of the label.
+        - (*property*) `canvas.{x, y}.label.visible`: the visibility of the label.
+    - `canvas.{x, y}.ticks`: the x/y-axis ticks.
+        - (*property*) `canvas.{x, y}.ticks.pos`: the position of the ticks (read-only).
+        - (*property*) `canvas.{x, y}.ticks.labels`: the labels of the ticks (read-only).
+        - (*property*) `canvas.{x, y}.ticks.color`: the color of the ticks.
+        - (*property*) `canvas.{x, y}.ticks.size`: the size of the ticks.
+        - (*property*) `canvas.{x, y}.ticks.rotation`: the color of the tick labels.
+        - (*property*) `canvas.{x, y}.ticks.visible`: the visibility of the ticks.
+
+- `canvas.title`: the title object of the canvas.
+    - (*property*) `canvas.title.text`: the text of the title.
+    - (*property*) `canvas.title.color`: the color of the title as `ndarray`.
+    - (*property*) `canvas.title.size`: the font size of the title.
+    - (*property*) `canvas.title.family`: the font family of the title.
+    - (*property*) `canvas.title.visible`: the visibility of the title.
+
+!!! note
+    `canvas.title`, `canvas.{x, y}.label` and `canvas.{x, y}.ticks` share the same
+    properties (`color`, `size`, `family` and `visible`) related to the text design.
+
+Event handlers are also organized in namespaces. At any level, the value-changed event
+for the parameter `X` is in `events.X`. For example, when the x-limit is changed, signal
+will be emitted from `canvas.x.events.lim`. See [Event Handling](../events/index.md) for
+more details.
+
+## Update Canvas Appearance
+
+Canvases have helper functions of name `update_*` to update the appearance. They return
+the canvas itself, so that you can chain the methods after the constructor.
+
+``` python
+# update the color of x/y-axis
+canvas = new_canvas().update_axes(color="gray")
+
+# update the text labels
+canvas = new_canvas().update_labels(x="time", y="value", title="My Plot")
+```
+
+## Add Layers
+
+All the plotting elements are added to the canvas as a "layer". In `whitecanvas`, a
+layer is rarely constructed directly, but is returned by the `add_*` methods. Same
+method **always** returns the same type of layer.
+
+``` python
+#!name: canvas_basics_0
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+line = canvas.add_line([0, 1, 2, 3], [0, 1, 1, 0])
+markers = canvas.add_markers([0, 1, 2, 3], [0, 1, 1, 0])
+canvas.show()
+```
+
+See [Layers](../layers/index.md) for more details.
diff --git a/docs/canvas/grid.md b/docs/canvas/grid.md
new file mode 100644
index 00000000..376b06bb
--- /dev/null
+++ b/docs/canvas/grid.md
@@ -0,0 +1,66 @@
+# Canvas Grid
+
+## Vertical/Horizontal Grid
+
+``` python
+#!name: canvas_grid_vertical
+from whitecanvas import vgrid
+
+canvas = vgrid(3, backend="matplotlib")
+
+c0 = canvas.add_canvas(0)
+c0.add_text(0, 0, "Canvas 0")
+c1 = canvas.add_canvas(1)
+c1.add_text(0, 0, "Canvas 1")
+c2 = canvas.add_canvas(2)
+c2.add_text(0, 0, "Canvas 2")
+canvas.show()
+```
+
+
+``` python
+#!name: canvas_grid_horizontal
+from whitecanvas import hgrid
+
+canvas = hgrid(3, backend="matplotlib")
+
+c0 = canvas.add_canvas(0)
+c0.add_text(0, 0, "Canvas 0")
+c1 = canvas.add_canvas(1)
+c1.add_text(0, 0, "Canvas 1")
+c2 = canvas.add_canvas(2)
+c2.add_text(0, 0, "Canvas 2")
+canvas.show()
+```
+
+## 2D Grid
+
+``` python
+#!name: canvas_grid_2d
+from whitecanvas import grid
+
+canvas = grid(2, 2, backend="matplotlib")
+
+for i, j in [(0, 0), (0, 1), (1, 0), (1, 1)]:
+    c = canvas.add_canvas(i, j)
+    c.add_text(0, 0, f"Canvas ({i}, {j})")
+canvas.show()
+```
+
+## Non-uniform Grid
+
+The `*_nonuniform` functions allow you to create a grid with non-uniform sizes.
+Instead of specifying the number of rows and columns, these functions take a list of size ratios.
+
+``` python
+#!name: canvas_grid_2d_nonuniform
+
+from whitecanvas import grid_nonuniform
+
+canvas = grid_nonuniform([1, 2], [2, 1], backend="matplotlib")
+
+for i, j in [(0, 0), (0, 1), (1, 0), (1, 1)]:
+    c = canvas.add_canvas(i, j)
+    c.add_text(0, 0, f"Canvas ({i}, {j})")
+canvas.show()
+```
diff --git a/docs/canvas/index.md b/docs/canvas/index.md
index d98db70d..13155336 100644
--- a/docs/canvas/index.md
+++ b/docs/canvas/index.md
@@ -1,106 +1,6 @@
 # Canvas
 
-In `whitecanvas`, a "canvas" is an object with a layer list, axes elements, labels and
-methods to handle them. A "grid" is a collection of canvases and is implemented with
-rendering backends.
-
-## Create Canvases
-
-Canvas is created by `new_canvas` method.
-
-- `backend`: the name of the backend to use.
-- `size`: the size of the canvas (width × height in pixels).
-- `palette`: the name of the color palette to use. Any input that is accepted by the
-  [cmap.Colormap](https://cmap-docs.readthedocs.io/en/latest/catalog/) can be used,
-  which includes:
-  1. The name of the built-in colormaps, such as "viridis", "plasma", etc.
-  2. Sequence of colors, such as `["red", "green", "blue"]`.
-  3. A `cmap.Colormap` object.
-
-``` python
-from whitecanvas import new_canvas
-
-canvas = new_canvas("matplotlib", size=(400, 300), palette="tab10")
-```
-
-A `SingleCanvas` object is returned. It can be considered as a mixture of a canvas
-and a grid. All of the functions mentioned above are implemented in this object.
-Therefore, you can show the canvas by calling the `show` method.
-
-``` python
-#!skip
-canvas.show()  # show the canvas, depending on the backend
-```
-
-## Namespaces
-
-To make the API tidy, each element of a canvas is organized in namespaces. By using
-namespaces, we can avoid the long list of arguments and make the code more readable.
-For example, the x-limit of the canvas is controlled by the property `canvas.x.lim`.
-
-Following list shows the namespaces and the properties associated with them.
-
-- `canvas.{x, y}`: the x/y-axis.
-    - (*property*) `canvas.{x, y}.lim`: the x/y-limit as a tuple of floats.
-    - (*property*) `canvas.{x, y}.color`: the color of the x/y-axis.
-    - (*property*) `canvas.{x, y}.flipped`: whether the x/y-axis is flipped.
-    - `canvas.{x, y}.label`: the x/y-axis label.
-        - (*property*) `canvas.{x, y}.label.text`: the text of the label.
-        - (*property*) `canvas.{x, y}.label.color`: the color of the label as `ndarray`.
-        - (*property*) `canvas.{x, y}.label.size`: the font size of the label.
-        - (*property*) `canvas.{x, y}.label.family`: the font family of the label.
-        - (*property*) `canvas.{x, y}.label.visible`: the visibility of the label.
-    - `canvas.{x, y}.ticks`: the x/y-axis ticks.
-        - (*property*) `canvas.{x, y}.ticks.pos`: the position of the ticks (read-only).
-        - (*property*) `canvas.{x, y}.ticks.labels`: the labels of the ticks (read-only).
-        - (*property*) `canvas.{x, y}.ticks.color`: the color of the ticks.
-        - (*property*) `canvas.{x, y}.ticks.size`: the size of the ticks.
-        - (*property*) `canvas.{x, y}.ticks.rotation`: the color of the tick labels.
-        - (*property*) `canvas.{x, y}.ticks.visible`: the visibility of the ticks.
-
-- `canvas.title`: the title object of the canvas.
-    - (*property*) `canvas.title.text`: the text of the title.
-    - (*property*) `canvas.title.color`: the color of the title as `ndarray`.
-    - (*property*) `canvas.title.size`: the font size of the title.
-    - (*property*) `canvas.title.family`: the font family of the title.
-    - (*property*) `canvas.title.visible`: the visibility of the title.
-
-!!! note
-    `canvas.title`, `canvas.{x, y}.label` and `canvas.{x, y}.ticks` share the same
-    properties (`color`, `size`, `family` and `visible`) related to the text design.
-
-Event handlers are also organized in namespaces. At any level, the value-changed event
-for the parameter `X` is in `events.X`. For example, when the x-limit is changed, signal
-will be emitted from `canvas.x.events.lim`. See [Event Handling](../events/index.md) for
-more details.
-
-## Update Canvas Appearance
-
-Canvases have helper functions of name `update_*` to update the appearance. They return
-the canvas itself, so that you can chain the methods after the constructor.
-
-``` python
-# update the color of x/y-axis
-canvas = new_canvas().update_axes(color="gray")
-
-# update the text labels
-canvas = new_canvas().update_labels(x="time", y="value", title="My Plot")
-```
-
-## Add Layers
-
-All the plotting elements are added to the canvas as a "layer". In `whitecanvas`, a
-layer is rarely constructed directly, but is returned by the `add_*` methods. Same
-method **always** returns the same type of layer.
-
-``` python
-#!name: canvas_basics_0
-from whitecanvas import new_canvas
-
-canvas = new_canvas("matplotlib")
-line = canvas.add_line([0, 1, 2, 3], [0, 1, 1, 0])
-markers = canvas.add_markers([0, 1, 2, 3], [0, 1, 1, 0])
-canvas.show()
-```
-
-See [Layers](../layers/index.md) for more details.
+- [Basics](basics.md)
+- [Namespaces](namespaces.md)
+- [Canvas Grid](grid.md)
+- [Working with the Backend Objects](native_objects.md)
diff --git a/docs/layers/face_layers.md b/docs/layers/face_layers.md
index 9fee11aa..0a09e89a 100644
--- a/docs/layers/face_layers.md
+++ b/docs/layers/face_layers.md
@@ -2,7 +2,6 @@
 
 There are several layers that is composed of faces and edges.
 
-- `Markers` ... a layers composed of markers for scatter plots.
 - `Bars` ... a layer composed of bars.
 - `Band` ... a layer composed of a band region (fill-between region).
 - `Spans` ... a layer composed of infinitely long spans.
diff --git a/docs/layers/index.md b/docs/layers/index.md
index e69de29b..b7c1fc87 100644
--- a/docs/layers/index.md
+++ b/docs/layers/index.md
@@ -0,0 +1,5 @@
+# Layers
+
+- [Lines](lines.md)
+- [Markers](markers.md)
+- [Layer Groups](layer_groups.md)
diff --git a/docs/layers/line_layers.md b/docs/layers/line_layers.md
deleted file mode 100644
index 8cd4f2c2..00000000
--- a/docs/layers/line_layers.md
+++ /dev/null
@@ -1,42 +0,0 @@
-# Line-type Layers
-
-There are several layers that is composed of only lines.
-
-- `Line` ... a simple line.
-- `InfLine` ... a straight line that extends to infinity
-- `InfCurve` ... a curve that extends to infinity
-- `Errorbar` ... lines representing error bars
-- `Rug` ... lines representing rug plots
-
-These layers have following properties in common.
-
--   `color` ... color of the lines. Any color-like object is accepted.
--   `width` ... width of the lines. Should be a non-negative number.
--   `style` ... style of the lines. Should be one of `"-"`, `":"`,
-    `"-."`, `"--"`.
-
-!!! note
-    `style` is not supported in some backends.
-
-These properties can be configured in function calls, via properties or the `update`
-method.
-
-``` python
-#!name: line_layers_properties
-import numpy as np
-from whitecanvas import new_canvas
-
-canvas = new_canvas("matplotlib")
-
-# function call
-layer = canvas.add_line(np.arange(10), color="black", width=2, style=":")
-
-# properties
-layer.color = "#FF36D9"
-layer.width = 2.5
-layer.style = "-"
-
-# update method
-layer.update(color=[0.0, 1.0, 0.0, 1.0], width=1, style="--")
-canvas.show()
-```
diff --git a/docs/layers/lines.md b/docs/layers/lines.md
new file mode 100644
index 00000000..46b327c8
--- /dev/null
+++ b/docs/layers/lines.md
@@ -0,0 +1,111 @@
+# Lines
+
+There are several layers that is composed of only lines.
+
+- `Line` ... a simple line.
+- `InfLine` ... a straight line that extends to infinity
+- `InfCurve` ... a curve that extends to infinity
+- `Errorbar` ... lines representing error bars
+- `Rug` ... lines representing rug plots
+
+These layers have following properties in common.
+
+- `color` ... color of the lines. Any color-like object is accepted.
+- `width` ... width of the lines. Should be a non-negative number.
+- `style` ... style of the lines. Should be one of `"-"`, `":"`, `"-."`, `"--"`.
+
+!!! note
+    `style` is not supported in some backends.
+
+These properties can be configured in function calls, via properties or the `update`
+method.
+
+``` python
+#!name: line_layer_properties
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+# function call
+layer = canvas.add_line([0, 2, 1, 3, 4, 2, -1], color="black", width=2, style=":")
+
+# properties
+layer.color = "#FF36D9"
+layer.width = 2.5
+layer.style = "-"
+
+# update method
+layer.update(color=[0.0, 1.0, 0.0, 1.0], width=1, style="--")
+canvas.show()
+```
+
+## Line
+
+`Line` is a simple line defined by two arrays of x and y coordinates. It is usually
+created by the [`add_line`][whitecanvas.canvas.CanvasBase.add_line] method. This method
+accepts several ways to define the line coordinates.
+
+``` python
+#!name: line_layer_coordinates
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+canvas.add_line([0, 1, 0, -1, 0])  # only y values
+canvas.add_line([0, 1, 2, 3, 4], [1, 2, 1, 0, 1])  # x and y values
+canvas.add_line(np.arange(5), np.array([2, 3, 2, 1, 2]))  # numpy arrays
+canvas.add_line(np.array([[0, 3], [1, 4], [2, 3], [3, 2], [4, 3]]))  # (N, 2) array
+canvas.show()
+```
+
+## InfLine
+
+`InfLine` is a straight line that extends to infinity. Practically, it is achieved by
+connecting a callback that updates the line coordinates when the canvas view range is
+updated.
+
+`InfLine` is usually created by the [`add_infline`][whitecanvas.canvas.CanvasBase.add_infline] method, or in the specific cases, [`add_vline`][whitecanvas.canvas.CanvasBase.add_vline] and [`add_hline`][whitecanvas.canvas.CanvasBase.add_hline] for vertical and horizontal lines, respectively.
+
+``` python
+#!name: infline_layer
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+canvas.add_infline((0, 1), 45, color="black")  # y = x + 1
+canvas.add_vline(2, color="red")  # x = 2
+canvas.add_hline(-1, color="blue")  # y = -1
+canvas.x.lim = (-3, 3)
+canvas.y.lim = (-3, 3)
+canvas.show()
+```
+
+## InfCurve
+
+`InfCurve` is a curve that extends to infinity, defined by an arbitrary model function.
+Practically, it is achieved by connecting a callback that resamples the curve when the
+canvas view range is updated.
+
+`InfCurve` is usually created by the [`add_infcurve`][whitecanvas.canvas.CanvasBase.
+add_infcurve] method. The input model function must be defined as `model(x, ...)` where
+`x` is `ndarray` of the x-coordinates and the rest of the arguments are the parameters
+of the model. The parameters can be set by the `with_params` method of the returned
+layer.
+
+``` python
+#!name: infcurve_layer
+import numpy as np
+from whitecanvas import new_canvas
+
+def model(x, freq, phase):
+    return np.sin(x * freq - phase)
+
+canvas = new_canvas("matplotlib")
+canvas.add_infcurve(model, color="black").with_params(freq=2, phase=0)
+canvas.add_infcurve(model, color="black", style=":").with_params(freq=2, phase=1.6)
+canvas.x.lim = (-3, 3)
+canvas.y.lim = (-3, 3)
+canvas.show()
+```
diff --git a/docs/layers/markers.md b/docs/layers/markers.md
new file mode 100644
index 00000000..8dc51c0d
--- /dev/null
+++ b/docs/layers/markers.md
@@ -0,0 +1,254 @@
+# Markers
+
+`Markers` is a layer for scatter plots. It can be created with the `add_markers` method.
+
+``` python
+#!name: markers_layer
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+layer = canvas.add_markers([0, 1, 2, 3], [0, 1, 1, 0])
+canvas.show()
+```
+
+Symbol and size of the markers can easily be configured with the `symbol` and `size`
+arguments.
+
+``` python
+#!name: markers_layer_with_symbol_and_size
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+for i, symbol in enumerate(["o", "s", "x", "+"]):
+    for j, size in enumerate([5, 10, 15, 20]):
+        x = [i - 0.2, i - 0.2, i + 0.2, i + 0.2]
+        y = [j - 0.2, j + 0.2, j - 0.2, j + 0.2]
+        layer = canvas.add_markers(x, y, symbol=symbol, size=size)
+canvas.show()
+```
+
+!!! note
+    If the symbol is edge-only, its color and visibility are automatically updated to
+    ensure the markers are informative.
+
+## Marker Properties
+
+`Markers` layer has two namespaces: `face` and `edge`. `face` has following properties:
+
+- `color` ... color of the faces. Any color-like object is accepted.
+- `hatch` ... hatch pattern of the faces. Should be one of `""`, `"-"`, `"|`, `"+"`,
+  `"/"`, `"\\"`, `"x"` or `"."`.
+
+!!! note
+    `hatch` is not supported in some backends.
+
+`edge` has following properties:
+
+- `color` ... color of the lines. Any color-like object is accepted.
+- `width` ... width of the lines. Should be a non-negative number.
+- `style` ... style of the lines. Should be one of `"-"`, `":"`, `"-."`, `"--"`.
+
+!!! note
+    `style` is not supported in some backends.
+
+Methods for adding these layers always configure the `face` properties with the
+arguments. You can use the `with_edge` method of the output layer to set edge
+properties. This separation is very helpful to prevent the confusion of the arguments,
+especially the colors.
+
+``` python
+#!name: markers_layer_with_edge
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+canvas.add_markers(np.sin(np.arange(10)), color="yellow").with_edge(color="black")
+canvas.show()
+```
+
+All the properties can be set via properties of `face` and `edge`, or the `update`
+method.
+
+``` python
+layer.face.color = "yellow"
+layer.face.hatch = "x"
+
+layer.edge.color = "black"
+layer.edge.width = 2
+layer.edge.style = "--"
+
+# use `update`
+layer.face.update(color="yellow", hatch="x")
+layer.edge.update(color="black", width=2, style="--")
+```
+
+## Multi-face and Multi-edge Markers
+
+`Markers` supports multi-face and multi-edge. This means that you can create a layer
+with multiple colors, widths, etc.
+
+To do this, you have to call `with_face_multi` or `with_edge_multi` method.
+Here's an example of `Markers` with multi-faces.
+
+``` python
+#!name: markers_layer_multifaces
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+layer = (
+    canvas
+    .add_markers(np.arange(10), np.sin(np.arange(10)))
+    .with_face_multi(color=np.random.random((10, 3)))  # random colors
+)
+canvas.show()
+```
+
+After calling `with_face_multi`, the layer `face` property will return arrays instead
+of scalar values.
+
+``` python
+#!skip
+layer.face.color  # (N, 4) array of RGBA colors
+layer.face.hatch  # (N,) array of hatchs
+layer.face.alpha # (N,) array of alpha values
+```
+
+!!! note
+    IDE **can** detect whether a `Markers` layer is multi-face or not. `Markers` class
+    is a generic class with type variables for the face types. Since `with_face_multi`
+    is correctly typed, IDE will recognize the returned layer as a multi-face layer.
+
+Similarly, you can use `with_edge_multi` to create multi-edge markers.
+
+``` python
+#!name: markers_layer_multiedges
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+layer = (
+    canvas
+    .add_markers(np.arange(10), np.sin(np.arange(10)))
+    .with_edge_multi(width=np.abs(np.sin(np.arange(10))) * 2.5)
+)
+canvas.show()
+```
+
+## Multi-size Markers
+
+`with_size_multi` method can be used to create multi-size markers.
+
+``` python
+#!name: markers_layer_multisizes
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+layer = (
+    canvas
+    .add_markers(np.arange(10), np.sin(np.arange(10)))
+    .with_size_multi(np.abs(np.sin(np.arange(10))) * 16)
+)
+canvas.show()
+```
+
+
+## Hover Text
+
+### Give a sequence of hover texts
+
+`with_hover_text` method sets the hover text of the markers.
+
+``` python
+#!skip
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+layer = (
+    canvas
+    .add_markers(np.arange(10), np.sin(np.arange(10)))
+    .with_hover_text([f"point {i}" for i in range(10)])
+)
+canvas.show()
+```
+
+### Give a hover template
+
+``` python
+#!skip
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+
+layer = (
+    canvas
+    .add_markers(np.arange(10), np.sin(np.arange(10)))
+    .with_hover_template("x={x}, y={y}, i={i}")
+)
+canvas.show()
+```
+
+## Picking Markers
+
+Markers have a `picked` event. You can connect a callback to the event to handle the
+picking.
+
+``` python
+#!skip
+import numpy as np
+from whitecanvas import new_canvas
+
+canvas = new_canvas("matplotlib")
+layer = canvas.add_markers([0, 1, 2], [0, 0, 0])
+
+@layer.events.picked.connect
+def _on_pick(picked):
+    print(f"picked indices: {picked}")
+```
+
+## Methods for Better Interpretability
+
+`Markers` is implemented with the following methods to make the plot more interpretable.
+
+### Color by density
+
+Scatter plot is good at visualizing the outliers, but not at telling the distribution
+when the density is high. In this case, `color_by_density` method is very useful. It
+colors the markers by the density of the points using kernel density estimation.
+
+``` python
+#!name: markers_layer_color_by_density
+#!width: 500
+import numpy as np
+from whitecanvas import hgrid
+
+rng = np.random.default_rng(999)
+x = rng.normal(size=1000)
+y = rng.normal(size=1000)
+
+canvas = hgrid(2, backend="matplotlib")
+(
+    canvas
+    .add_canvas(0)
+    .update_labels(title="no coloring")
+    .add_markers(x, y)
+)
+(
+    canvas
+    .add_canvas(1)
+    .update_labels(title="with coloring")
+    .add_markers(x, y)
+    .color_by_density(cmap="viridis")
+)
+canvas.show()
+```
diff --git a/mkdocs.yml b/mkdocs.yml
index 7efa2d17..963bd375 100644
--- a/mkdocs.yml
+++ b/mkdocs.yml
@@ -24,11 +24,14 @@ nav:
   - Quick Start: quick_start.md
   - Canvas:
     - Overview: canvas/index.md
+    - Basics: canvas/basics.md
     - Namespaces: canvas/namespaces.md
+    - Canvas Grid: canvas/grid.md
     - Working with the Backend Objects: canvas/native_objects.md
   - Layers:
     - Overview: layers/index.md
-    - Line Layers: layers/line_layers.md
+    - Lines: layers/lines.md
+    - Markers: layers/markers.md
     - Face&Edge Layers: layers/face_layers.md
     - Layer Groups: layers/layer_groups.md
   - Categorical Plot:
diff --git a/whitecanvas/canvas/_base.py b/whitecanvas/canvas/_base.py
index 8327c5e0..15ea8a4e 100644
--- a/whitecanvas/canvas/_base.py
+++ b/whitecanvas/canvas/_base.py
@@ -352,13 +352,16 @@ def cat(
         data : tabular data
             Any categorizable data. Currently, dict, pandas.DataFrame, and
             polars.DataFrame are supported.
-
+        x : str, optional
+            Name of the column that will be used for the x-axis. Must be numerical.
+        y : str, optional
+            Name of the column that will be used for the y-axis. Must be numerical.
         update_labels : bool, default True
             If True, update the x/y labels to the corresponding names.
 
         Returns
         -------
-        CategorizedPlot
+        CatPlotter
             Plotter object.
         """
         plotter = _df.CatPlotter(self, data, x, y, update_label=update_labels)
@@ -372,6 +375,26 @@ def cat_x(
         *,
         update_labels: bool = True,
     ) -> _df.XCatPlotter[Self, _DF]:
+        """
+        Categorize input data for plotting with x-axis as a categorical axis.
+
+        Parameters
+        ----------
+        data : tabular data
+            Any categorizable data. Currently, dict, pandas.DataFrame, and
+            polars.DataFrame are supported.
+        x : str or sequence of str, optional
+            Name of the column(s) that will be used for the x-axis. Must be categorical.
+        y : str, optional
+            Name of the column that will be used for the y-axis. Must be numerical.
+        update_labels : bool, default True
+            If True, update the x/y labels to the corresponding names.
+
+        Returns
+        -------
+        XCatPlotter
+            Plotter object.
+        """
         return _df.XCatPlotter(self, data, x, y, update_labels)
 
     def cat_y(
@@ -382,6 +405,26 @@ def cat_y(
         *,
         update_labels: bool = True,
     ) -> _df.YCatPlotter[Self, _DF]:
+        """
+        Categorize input data for plotting with y-axis as a categorical axis.
+
+        Parameters
+        ----------
+        data : tabular data
+            Any categorizable data. Currently, dict, pandas.DataFrame, and
+            polars.DataFrame are supported.
+        x : str, optional
+            Name of the column that will be used for the x-axis. Must be numerical.
+        y : str or sequence of str, optional
+            Name of the column(s) that will be used for the y-axis. Must be categorical.
+        update_labels : bool, default True
+            If True, update the x/y labels to the corresponding names.
+
+        Returns
+        -------
+        YCatPlotter
+            Plotter object
+        """
         return _df.YCatPlotter(self, data, y, x, update_labels)
 
     def cat_xy(
@@ -392,6 +435,26 @@ def cat_xy(
         *,
         update_labels: bool = True,
     ) -> _df.XYCatPlotter[Self, _DF]:
+        """
+        Categorize input data for plotting with both axes as categorical.
+
+        Parameters
+        ----------
+        data : tabular data
+            Any categorizable data. Currently, dict, pandas.DataFrame, and
+            polars.DataFrame are supported.
+        x : str or sequence of str, optional
+            Name of the column(s) that will be used for the x-axis. Must be categorical.
+        y : str or sequence of str, optional
+            Name of the column(s) that will be used for the y-axis. Must be categorical.
+        update_labels : bool, default True
+            If True, update the x/y labels to the corresponding names.
+
+        Returns
+        -------
+        XYCatPlotter
+            Plotter object
+        """
         return _df.XYCatPlotter(self, data, x, y, update_labels)
 
     def stack_over(self, layer: _L0) -> StackOverPlotter[Self, _L0]:
@@ -406,11 +469,13 @@ def stack_over(self, layer: _L0) -> StackOverPlotter[Self, _L0]:
 
         will result in a bar plot like this
 
+        ```
          ┌───┐
          ├───│┌───┐
          │   │├───│
          ├───│├───│
         ─┴───┴┴───┴─
+        ```
         """
         if not isinstance(layer, (_l.Bars, _l.Band, _lg.StemPlot, _lg.LabeledBars)):
             raise TypeError(
@@ -968,7 +1033,7 @@ def add_infcurve(
         width = theme._default("line.width", width)
         style = theme._default("line.style", style)
         layer = _l.InfCurve(
-            model, bounds=bounds, name=name, color=color, width=width,
+            model, bounds=bounds, name=name, color=color, width=width, alpha=alpha,
             style=style, antialias=antialias, backend=self._get_backend(),
         )  # fmt: skip
         return self.add_layer(layer)
@@ -979,11 +1044,36 @@ def add_hline(
         *,
         name: str | None = None,
         color: ColorType | None = None,
-        width: float = 1.0,
+        width: float | None = None,
         style: LineStyle | str = LineStyle.SOLID,
         alpha: float = 1.0,
         antialias: bool = True,
     ) -> _l.InfLine:
+        """
+        Add a infinite horizontal line to the canvas.
+
+        Parameters
+        ----------
+        y : float
+            Y coordinate of the line.
+        name : str, optional
+            Name of the layer.
+        color : color-like, optional
+            Color of the bars.
+        width : float, optional
+            Line width. Use the theme default if not specified.
+        style : str or LineStyle, optional
+            Line style. Use the theme default if not specified.
+        alpha : float, default 1.0
+            Alpha channel of the line.
+        antialias : bool, default True
+            Antialiasing of the line.
+
+        Returns
+        -------
+        InfLine
+            The infline layer.
+        """
         return self.add_infline(
             (0, y), 0, name=name, color=color, width=width, style=style, alpha=alpha,
             antialias=antialias
@@ -995,11 +1085,36 @@ def add_vline(
         *,
         name: str | None = None,
         color: ColorType | None = None,
-        width: float = 1.0,
+        width: float | None = None,
         style: LineStyle | str = LineStyle.SOLID,
         alpha: float = 1.0,
         antialias: bool = True,
     ) -> _l.InfLine:
+        """
+        Add a infinite vertical line to the canvas.
+
+        Parameters
+        ----------
+        x : float
+            X coordinate of the line.
+        name : str, optional
+            Name of the layer.
+        color : color-like, optional
+            Color of the bars.
+        width : float, optional
+            Line width. Use the theme default if not specified.
+        style : str or LineStyle, optional
+            Line style. Use the theme default if not specified.
+        alpha : float, default 1.0
+            Alpha channel of the line.
+        antialias : bool, default True
+            Antialiasing of the line.
+
+        Returns
+        -------
+        InfLine
+            The infline layer.
+        """
         return self.add_infline(
             (x, 0), 90, name=name, color=color, width=width, style=style, alpha=alpha,
             antialias=antialias,