The right way to Use Conditional Formatting in Pandas to Improve Knowledge Visualization

df.model.<technique>(<arguments>)

The place <technique> is the precise formatting perform you need to apply, and <arguments> are the parameters required by that perform. The Styler object returns the formatted dataframe with out altering the unique one. There are two approaches to utilizing conditional formatting with the Styler object:

• Constructed-in Types: To use fast formatting types to your dataframe
• Customized Stylization: Create your individual formatting guidelines for the Styler object and cross them by one of many following strategies (Styler.applymap: element-wise or Styler.apply: column-/row-/table-wise)

Now, we'll cowl some examples of each approaches that will help you improve the visualization of your knowledge.

Examples: Constructed-in-Types

Let’s create a dummy inventory value dataset with columns for Date, Value Value, Satisfaction Rating, and Gross sales Quantity to display the examples under:

import pandas as pd
import numpy as np

knowledge = {'Date': ['2024-03-05', '2024-03-06', '2024-03-07', '2024-03-08', '2024-03-09', '2024-03-10'],
        'Value Value': [100, 120, 110, 1500, 1600, 1550],
        'Satisfaction Rating': [90, 80, 70, 95, 85, 75],
        'Gross sales Quantity': [1000, 800, 1200, 900, 1100, None]}

df = pd.DataFrame(knowledge)
df

Output:

Authentic Unformatted Dataframe

1. Highlighting Most and Minimal Values

We will use highlight_max and highlight_min capabilities to spotlight the utmost and minimal values in a column or row. For column set axis=0 like this:

# Highlighting Most and Minimal Values
df.model.highlight_max(shade="inexperienced", axis=0 , subset=['Cost Price', 'Satisfaction Score', 'Sales Amount']).highlight_min(shade="crimson", axis=0 , subset=['Cost Price', 'Satisfaction Score', 'Sales Amount'])

Output:
 

Max & Min Values

2. Making use of Shade Gradients

Shade gradients are an efficient solution to visualize the values in your knowledge. On this case, we'll apply the gradient to satisfaction scores utilizing the colormap set to 'viridis'. This can be a sort of shade coding that ranges from purple (low values) to yellow (excessive values). Right here is how you are able to do this:

# Making use of Shade Gradients
df.model.background_gradient(cmap='viridis', subset=['Satisfaction Score'])

Output:

Colormap - viridis

3. Highlighting Null or Lacking Values

When we've got massive datasets, it turns into tough to determine null or lacking values. You need to use conditional formatting utilizing the built-in df.model.highlight_null perform for this function. For instance, on this case, the gross sales quantity of the sixth entry is lacking. You may spotlight this data like this:

# Highlighting Null or Lacking Values
df.model.highlight_null('yellow', subset=['Sales Amount'])

Output:
 

Highlighting Lacking Values

Examples: Customized Stylization Utilizing apply() & applymap()

1.  Conditional Formatting for Outliers

Suppose that we've got a housing dataset with their costs, and we need to spotlight the homes with outlier costs (i.e., costs which are considerably increased or decrease than the opposite neighborhoods). This may be finished as follows:

import pandas as pd
import numpy as np

# Home costs dataset
df = pd.DataFrame({
   'Neighborhood': ['H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'H7'],
   'Value': [50, 300, 360, 390, 420, 450, 1000],
})

# Calculate Q1 (twenty fifth percentile), Q3 (seventy fifth percentile) and Interquartile Vary (IQR)
q1 = df['Price'].quantile(0.25)
q3 = df['Price'].quantile(0.75)
iqr = q3 - q1

# Bounds for outliers
lower_bound = q1 - 1.5 * iqr
upper_bound = q3 + 1.5 * iqr

# Customized perform to spotlight outliers
def highlight_outliers(val):
   if val < lower_bound or val > upper_bound:
      return 'background-color: yellow; font-weight: daring; shade: black'
   else:
      return ''

df.model.applymap(highlight_outliers, subset=['Price'])

Output:

Highlighting Outliers

2. Highlighting Tendencies

Think about that you simply run an organization and are recording your gross sales day by day. To research the tendencies, you need to spotlight the times when your day by day gross sales enhance by 5% or extra. You may obtain this utilizing a customized perform and the apply technique in pandas. Right here’s how:

import pandas as pd

# Dataset of Firm's Gross sales
knowledge = {'date': ['2024-02-10', '2024-02-11', '2024-02-12', '2024-02-13', '2024-02-14'],
        'gross sales': [100, 105, 110, 115, 125]}

df = pd.DataFrame(knowledge)

# Every day proportion change
df['pct_change'] = df['sales'].pct_change() * 100

# Spotlight the day if gross sales elevated by greater than 5%
def highlight_trend(row):
    return ['background-color: green; border: 2px solid black; font-weight: bold' if row['pct_change'] > 5 else '' for _ in row]

df.model.apply(highlight_trend, axis=1)

Output:

3. Highlighting Correlated Columns

Correlated columns are vital as a result of they present relationships between totally different variables. For instance, if we've got a dataset containing age, revenue, and spending habits and our evaluation exhibits a excessive correlation (near 1) between age and revenue, then it means that older individuals usually have increased incomes. Highlighting correlated columns helps to visually determine these relationships. This method turns into extraordinarily useful because the dimensionality of your knowledge will increase. Let's discover an instance to higher perceive this idea:

import pandas as pd

# Dataset of individuals
knowledge = {
    'age': [30, 35, 40, 45, 50],
    'revenue': [60000, 66000, 70000, 75000, 100000],
    'spending': [10000, 15000, 20000, 18000, 12000]
}

df = pd.DataFrame(knowledge)

# Calculate the correlation matrix
corr_matrix = df.corr()

# Spotlight extremely correlated columns
def highlight_corr(val):
    if val != 1.0 and abs(val) > 0.5:   # Exclude self-correlation
        return 'background-color: blue; text-decoration: underline'
    else:
        return ''

corr_matrix.model.applymap(highlight_corr)

Output:

Correlated Columns

Wrapping Up

These are simply a few of the examples I confirmed as a starter to up your recreation of knowledge visualization. You may apply related methods to varied different issues to reinforce the information visualization, comparable to highlighting duplicate rows, grouping into classes and deciding on totally different formatting for every class, or highlighting peak values. Moreover, there are numerous different CSS choices you may discover within the official documentation. You may even outline totally different properties on hover, like magnifying textual content or altering shade. Take a look at the "Enjoyable Stuff" part for extra cool concepts. This text is a part of my Pandas sequence, so in the event you loved this, there's loads extra to discover. Head over to my creator web page for extra ideas, tips, and tutorials.

Kanwal Mehreen Kanwal is a machine studying engineer and a technical author with a profound ardour for knowledge science and the intersection of AI with medication. She co-authored the book "Maximizing Productiveness with ChatGPT". As a Google Era Scholar 2022 for APAC, she champions variety and educational excellence. She's additionally acknowledged as a Teradata Range in Tech Scholar, Mitacs Globalink Analysis Scholar, and Harvard WeCode Scholar. Kanwal is an ardent advocate for change, having based FEMCodes to empower ladies in STEM fields.