Seasonal Decompose Plot

General

Let’s begin by importing the dataset module of vastorbit. It provides a range of datasets for both training and exploring vastorbit’s capabilities.

import vastorbit as vo
import vastorbit.datasets as vod

Let’s utilize the Airline Passenger dataset to demonstrate time series capabilities.

passengers = vod.load_airline_passengers()

This dataset is well-suited for seasonal decomposition. It represents the time series of the number of passengers for a specific flight since 1950. Notably, it exhibits a noticeable trend and seasonality pattern. It is evident that the time series follows a multiplicative model.

Let’s perform a decomposition of the time series.

But before that let’s import the vastorbit statistical tests and decompose the time series.

import vastorbit.machine_learning.model_selection.statistical_tests as vmlt

# We use seasonal decompose to get the time series component
decomposition = vms.seasonal_decompose(
    input_relation = passengers,
    columns = "passengers",
    ts = "date",
    polynomial_order = 2,
    mult = True,
    use_row = False,
)

To create a seasonal decomposition plot, we must visualize the primary time series along with all its individual components. This decomposition process extracts various time series components, and we can then proceed to visualize each of them separately.

Plotly

We can switch to using the plotly module.

vo.set_option("plotting_lib", "plotly")

We can first create custom subplot array using plotly. The dimensions are set to 3 x 2.

from plotly.subplots import make_subplots

fig = make_subplots(rows=3, cols=2, column_widths=[0.7, 0.3],
                    specs=[[{"rowspan": 3}, {"rowspan": 1}],
                            [{}, {"rowspan": 1}],
                            [{}, {"rowspan": 1}]],)

Then we can indivually add the traces.

# Add the first trace (spans three rows)
fig.add_trace(
    decomposition["passengers"].plot(ts = "date", colors = "#0073E7").data[0],
    row=1, col=1,
)

# Add the second trace (second column, first row)
fig.add_trace(
    decomposition["passengers_trend"].plot(ts = "date", colors = "black").data[0],
    row=1, col=2
)

# Add the third trace (second column, second row)
fig.add_trace(
    decomposition["passengers_seasonal"].plot(ts = "date", colors = "green").data[0],
    row=2, col=2
)

# Add the fourth trace (third row, second column)
fig.add_trace(
    decomposition["passengers_epsilon"].plot(ts = "date",colors = "grey").data[0],
    row=3, col=2
)
fig.update_layout(height = 500, width = 700)

Matplotlib

We can switch to using the matplotlib module.

vo.set_option("plotting_lib", "matplotlib")

We need to import matplotlib.pyplot.

import matplotlib.pyplot as plt

We’ll create four subplots for this purpose.

fig = plt.figure()
fig.set_size_inches(10, 6)

ax1 = fig.add_subplot(121)
ax2 = fig.add_subplot(322)
ax3 = fig.add_subplot(324)
ax4 = fig.add_subplot(326)

Following this, we can proceed to plot the final visualization.

decomposition["passengers"].plot(ts = "date", ax = ax1, color = "#0073E7")
decomposition["passengers_trend"].plot(ts = "date", ax = ax2, color = "black")
ax2.set_xlabel("")
ax2.get_xaxis().set_ticks([])
decomposition["passengers_seasonal"].plot(ts = "date", ax = ax3, color = "green")
ax3.set_xlabel("")
ax3.get_xaxis().set_ticks([])

decomposition["passengers_epsilon"].plot(ts = "date", ax = ax4, color = "grey")

---

Chart Customization

vastorbit empowers users with a high degree of flexibility when it comes to tailoring the visual aspects of their plots. This customization extends to essential elements such as color schemes, text labels, and plot sizes, as well as a wide range of other attributes that can be fine-tuned to align with specific design preferences and analytical requirements. Whether you want to make your visualizations more visually appealing or need to convey specific insights with precision, vastorbit’s customization options enable you to craft graphics that suit your exact needs.

Note

As seasonal decomposition plots consist of multiple line charts, we recommend referring to the page on customizing Line Plots charts for guidance on customization.