Introduction: Real World Example

Now that you get a general idea of how to create composable visualization in Marsilea. Let’s apply it on a real dataset.

Load Dataset

We use a cooking oils dataset. This dataset contains the fat content in cooking oils.

>>> import marsilea as ma
>>> import marsilea.plotter as mp
>>> data = ma.load_data('cooking_oils')
>>> fat_content = data[['saturated', 'polyunsaturated (omega 3 & 6)',
...                     'monounsaturated', 'other fat']]

The main feature is the fat content. Let’s visualize it with stack bar plot.

Create main canvas

Like what we’ve introduced in the previous section, the main canvas must be created first. Here, we create an empty ClusterBoard.

Define some colors and create a stack bar plot.

>>> red, yellow, green, gray, blue = "#cd442a", "#f0bd00", "#7e9437", "#eee", "#009de1"
>>> sb = mp.StackBar(fat_content.T * 100, colors=[red, yellow, green, gray],
...                  width=.8, orient="h", label="Fat Content (%)",
...                  legend_kws={'ncol': 2, 'fontsize': 10})

Let’s create a canvas to draw the stack bar.

>>> cb = ma.ClusterBoard(fat_content.to_numpy(), height=10, margin=.5)
>>> cb.add_layer(sb)
>>> cb.render()

We use the add_layer() to add the stack bar plot to the canvas.

Add side plots

>>> oil_name = mp.Labels(data.index.str.capitalize())
>>> fmt = lambda x: f"{x:.1f}" if x > 0 else ""
>>> trans_fat = mp.Numbers(data['trans fat'] * 100, label="Trans Fat (%)",
...                        fmt=fmt, color=blue)
>>> cb.add_right(oil_name, size=1.1, pad=.2)
>>> cb.add_left(trans_fat, pad=.2, name="trans fat")
>>> cb.render()

Here we use add_right() and add_left() method to add texts and bar plot to the right and left side of the stack bar plot. Adjustment of plot size and space between plots is achieved with size and pad parameter.

The trans_fat is named with name parameter. This name will be used later to retrieve the axes.

Grouping

We split the data into groups based on the cooking conditions.

>>> order=["Control", ">230 °C (Deep-frying)", "200-229 °C (Stir-frying)",
...        "150-199 °C (Light saute)", "<150 °C (Dressings)"]
>>> group_color = ["#e5e7eb", "#c2410c", "#fb923c", "#fca5a5", "#fecaca"]
>>> cb.group_rows(data['cooking conditions'], order=order)

Add text to annotate the groups.

>>> chunk_text=["Control", ">230 °C\nDeep-frying", "200-229 °C\nStir-frying",
...             "150-199 °C\nLight saute", "<150 °C\nDressings"]
>>> cb.add_left(mp.Chunk(chunk_text, group_color, rotation=0, padding=10), pad=.1)
>>> cb.render()

Add dendrogram

Perform hierarchical clustering on the cooking oils within each group.

>>> cb.add_dendrogram("left", add_meta=False, colors=group_color,
...                   linewidth=1.5, size=.5, pad=.02)
>>> cb.render()

Add legends

>>> cb.add_legends("bottom", pad=.3)
>>> cb.render()

Here we add the legends to the bottom of the canvas using add_legends().

Make adjustment

Notice that the text render in the Trans Fat (%) bar plot is outside of the plot. We can adjust the x-axis to make it fit.

Remember that we named this plot with name before. We can retrieve the axes where the plot is drawn using get_ax().

Note

If you don’t know about what axes is, you can read this.

>>> cb.render()
>>> axes = cb.get_ax("trans fat")
>>> for ax in axes:
...     ax.set_xlim(4.2, 0)

The get_ax() will return a list of axes if we group the data. Otherwise, it will return a single axes.