JIYIK CN >

This article explores the concept of plotting series on a DataFrame using Pandas.

Whether you are exploring a dataset to hone your skills or aiming to make a good presentation for your company’s performance analysis, visualization plays an important role.

Python provides various options through its .plot()Pandas library with functions to transform our data into presentable forms in unprecedented ways.

Even an amateur Python developer can easily figure out how to use the library after understanding the steps and following the correct procedures to generate valuable insights.

However, to do this, we first need to understand what the library does and how it helps analysts provide value to their companies.

Various Plots in Pandas

Let's start this tutorial by understanding how many different graphs there are.

1. line- Line graph (this is the default graph)
2. bar- Bar chart parallel to the Y axis (vertical)
3. barh- Bar chart parallel to the X-axis (horizontal)
4. hist- Histogram
5. box- Box plot
6. kde- Kernel density estimation plot
7. density- kdeSame as
8. area- Area Chart
9. pie- Pie Chart

Pandas uses plot()the method for visualization. Additionally, the library pyplotcan be used Matplotlibfor illustrations.

This tutorial covers important plot types and how to use them effectively.

Plotting a Bar Chart from a Pandas Series

As the name suggests, series plots are important when the data is in the form of series, and there should be a correlation between the variables. If there is no correlation, we will not be able to visualize and compare.

Below is an example of drawing a basic bar chart based on dummy data given in dictionary form. We can use a CSV file based on real data or we can use custom created dummy data to explore various options for development and research.

import pandas as pd
import matplotlib.pyplot as plt

s = pd.Series(
    {
        16976: 2,
        1: 39,
        2: 49,
        3: 187,
        4: 159,
        5: 158,
        16947: 14,
        16977: 1,
        16948: 7,
        16978: 1,
        16980: 1,
    },
    name="article_id",
)
print(s)

# Name: article_id, dtype: int64
s.plot.bar()
plt.show()

The above code gives this output.

As we can see, a bar graph is displayed to aid in comparison and analysis.

Plotting a Line Chart of a Pandas Series

Let's consider one more example where our purpose is to draw a line chart based on given dummy data. Here, we should not add extra elements and plot().

# using Series.plot() method
s = pd.Series([0.1, 0.4, 0.16, 0.3, 0.9, 0.81])
s.plot()
plt.show()

The above code gives this output.

It is also possible to plot a graph with multiple variables on the Y-axis as shown below. Including multiple variables in a single graph makes it more illustrative and feasible to compare elements belonging to the same category.

For example, if a graph of students' scores in a particular exam is created, it will help the professor analyze the performance of each student in a specific time interval.

import numpy as np

ts = pd.Series(np.random.randn(1000), index=pd.date_range("1/1/2000", periods=1000))
df = pd.DataFrame(np.random.randn(1000, 4), index=ts.index, columns=list("ABCD"))
df = df.cumsum()
plt.figure()
df.plot()
plt.show()

Plotting a Boxplot of a Pandas Series

plot()The method allows for other plotting styles besides the default line plot. We can provide kindthe -p argument to the plot function.

We can series.plot.box()draw a box plot to illustrate the distribution of values ​​within each column by calling the function . A box plot tells us a lot about the data, such as the median.

We can also find out the first, second, and third quartiles by looking at the box plot.

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

df = pd.DataFrame(np.random.rand(10, 5), columns=["A", "B", "C", "D", "E"])
df.plot.box()
plt.show()

Additionally, colorpassing arguments of other types using the keyword will be immediately assigned to matplotlibfor all boxes, whiskers, , mediansand capsshadings.

We can get this infographic given below by writing just one line.

Draw a Histogram Bin Chart of a Pandas Series

Next, we will learn how to plot hexagonal bin plots and autocorrelation plots.

Histogram Bin plots are dataframe.plot.hexbin()created using the syntax. These are a good alternative to scatter plots if your data is too dense to plot each point clearly.

A very important keyword here is gridsize, as it controls the number of hexagons along the horizontal direction. More grids will tend towards smaller and larger bins.

Here is the following code snippet based on random data.

df = pd.DataFrame(np.random.randn(1000, 2), columns=["a", "b"])
df["b"] = df["b"] + np.arange(1000)
df["z"] = np.random.uniform(0, 3, 1000)
df.plot.hexbin(x="a", y="b", C="z", reduce_C_function=np.max, gridsize=25)
plt.show()

The above code gives this output.

For more information on hexagonal bin plots, navigate to the hexagonal bin plot method in the Pandas official documentation hexbin.

Plotting the Autocorrelation of a Pandas Series

We end this tutorial with the most complex type of plot: the autocorrelation plot. This plot is often used to analyze machine learning models based on neural networks.

It is used to describe whether the elements in a time series are positively correlated, negatively correlated, or not dependent on each other. We can find 自相关the value of the function ACF on the Y-axis, ranging from -1 to 1

It helps to correct for randomness in the time series. We obtain the data by calculating the autocorrelation at different time lags.

The lines parallel to the x-axis correspond to the approximately 95% to 99% confidence bands. The dashed lines are the 99% confidence bands.

Let's see how to create this graph.

from pandas.plotting import autocorrelation_plot

plt.figure()
spacing = np.linspace(-9 * np.pi, 9 * np.pi, num=1000)
data = pd.Series(0.7 * np.random.rand(1000) + 0.3 * np.sin(spacing))
autocorrelation_plot(data)
plt.show()

If the time series is not based on real data, then this autocorrelation is around zero for all lag differences, and if the time series is based on real data, then the autocorrelation must be non-zero. There must be one or more autocorrelations.