DEV Community

Cover image for Pandas Core Data Preprocessing Techniques - A Recap
Uche Emmanuel
Uche Emmanuel

Posted on

Pandas Core Data Preprocessing Techniques - A Recap

Introduction

Data preprocessing is a crucial step in any data analysis/data science/machine learning project. It involves transforming and cleaning raw data into a format that can be easily analyzed and visualized and used for modeling. In this document, I have provided a recap of some core data preprocessing techniques and procedures using Python's Pandas library.

Pandas is a popular open-source data analysis library for Python. It provides powerful data structures for working with structured data and a wide range of tools for data manipulation, analysis, and visualization.

Getting started

Before we dive into data preprocessing techniques, let's first ensure that Pandas is installed in your environment. You can install Pandas using pip:

!pip install pandas
Enter fullscreen mode Exit fullscreen mode

Once Pandas is installed, you can import it into your Python script or notebook using:

import pandas as pd
Enter fullscreen mode Exit fullscreen mode

Data loading and exploration

The first step in any data analysis/data science/machine learning project is to load the data and explore its structure and properties. Pandas provides several methods for loading data from various file formats, including CSV, Excel, SQL databases, and more.

Here's an example of loading a CSV file using Pandas:

data = pd.read_csv('data.csv')
Enter fullscreen mode Exit fullscreen mode

Here you have to ensure that the data file is in the current directory where you are writing your code. Once the data is loaded, we can explore its structure using methods such as head, tail, info, describe, and more. These methods provide useful information about the data, such as column names, data types, summary statistics, and sample rows.

# show the first five rows of the data
print(data.head())

# show the last five rows of the data
print(data.tail())

# show information about the data
print(data.info())

# show summary statistics of the data
print(data.describe())
Enter fullscreen mode Exit fullscreen mode

Data cleaning

After exploring the data, the next step is to clean it by handling missing values, duplicate data, outliers, and incorrect data types. Pandas provides several methods for handling these issues.

Handling missing values

Missing values are common in real-world datasets and can be problematic for data analysis. Pandas provides several methods for handling missing values, including dropna, fillna, and more.

Here's an example of dropping rows with missing values:

# drop rows with missing values
clean_data = data.dropna()
Enter fullscreen mode Exit fullscreen mode

And here's an example of filling missing values with a specific value:

# fill missing values with zero
clean_data = data.fillna(0)
Enter fullscreen mode Exit fullscreen mode

Handling duplicate data

Duplicate data can skew analysis results and should be removed before analysis. Pandas provides a drop_duplicates method for removing duplicate rows.

# drop duplicate rows
clean_data = data.drop_duplicates()
Enter fullscreen mode Exit fullscreen mode

Handling outliers

Outliers can also skew analysis results and should be handled appropriately. Pandas provides several methods for handling outliers, including clip and quantile.

Here's an example of clipping values at a certain threshold:

# clip values at 5th and 95th percentile
clean_data = data.clip(lower=data.quantile(0.05), upper=data.quantile(0.95), axis=1)
Enter fullscreen mode Exit fullscreen mode

Data transformation

After cleaning the data, the next step is to transform it into a format that can be easily analyzed and visualized. Pandas provides several methods for transforming data, including groupby, pivot_table, merge, and more.

Grouping data

Grouping data is a common operation in data analysis, and Pandas provides a groupby method for this purpose. Here's an
example of grouping data by a specific column and calculating the mean:

# group data by 'category' column and calculate the mean of 'value' column
grouped_data = data.groupby('category')['value'].mean()
Enter fullscreen mode Exit fullscreen mode

Pivoting data

Pivoting data involves reshaping data from a long format to a wide format. Pandas provides a pivot_table method for pivoting data.

Here's an example of pivoting data based on two columns:

# pivot data based on 'category' and 'date' columns
pivoted_data = pd.pivot_table(data, values='value', index='category', columns='date')
Enter fullscreen mode Exit fullscreen mode

Merging data

Merging data involves combining data from multiple sources based on a common column. Pandas provides a merge method for merging data.

Here's an example of merging two data frames based on a common column:

# merge two dataframes based on 'id' column
merged_data = pd.merge(df1, df2, on='id')
Enter fullscreen mode Exit fullscreen mode

Conclusion

In this dev post, I have provided a recap of some core data preprocessing techniques and procedures using Python's Pandas library. I covered data loading and exploration, data cleaning, and data transformation using methods such as dropna, fillna, drop_duplicates, groupby, pivot_table, and merge.

These are just some of the many techniques and procedures available in Pandas for data preprocessing. By mastering these techniques and combining them with other tools in your data analysis toolkit, you'll be well on your way to becoming a proficient data analyst or data scientist.

I hope that this dev post has been helpful in expanding your knowledge of Pandas and data preprocessing, and I wish you the best of luck in your future data projects!

Top comments (0)