Pandas Library in Python

Introduction

• Pandas Library: Essential for data science in Python
• Video Scope: From beginner to comfortable with Pandas
• Two Audiences: Newcomers and those looking for specific tips
• Flexibility: More than Excel, large datasets
• Install Pandas: pip install pandas

Loading Data

Steps

1. CSV Data: Download from GitHub (Example: Pokemon data)
   • Save as .csv
2. Text Editor: Open in Jupyter Notebooks recommended
3. Import Pandas: import pandas as pd
4. Load Data: df = pd.read_csv('pokemon_data.csv')

Data Inspection

• Print Head/Tail: df.head(3), df.tail(3)

Loading Different File Types

• Excel: df = pd.read_excel('pokemon_data.xlsx')
• Tab-Separated Values (TSV): df = pd.read_csv('pokemon_data.txt', delimiter='\t')

Data Manipulation Basics

Columns and Rows

• View Columns: df.columns
• Access Column: df['Name'] or df.Name
• Multiple Columns: df[['Name', 'Type 1', 'HP']]

Specific Rows and Values

• Single Row: df.iloc[1]
• Multiple Rows: df.iloc[1:4]
• Specific Value: df.iloc[2, 1]

Iterating Through Rows

• Iterate: for index, row in df.iterrows():
• Access Data: row['Name']

Conditional Selection

• Single Condition: df.loc[df['Type 1'] == 'Fire']
• Multiple Conditions: df.loc[(df['Type 1'] == 'Fire') & (df['HP'] > 70)]

Descriptive Statistics

• Describe Method: df.describe()
• Sort Values: df.sort_values('Name', ascending=False)

Advanced Data Manipulation

Adding/Dropping Columns

• Add Column: df['Total'] = df.iloc[:, 4:10].sum(axis=1)
• Drop Column: df.drop(columns=['Total'], inplace=True)

Saving Data

• To CSV: df.to_csv('modified.csv', index=False)
• To Excel: df.to_excel('modified.xlsx', index=False)
• To TSV: df.to_csv('modified.txt', sep='\t', index=False)

Advanced Filtering

• Regular Expressions: df.loc[df['Name'].str.contains('^Pi[a-z]*', flags=re.I, regex=True)]

Modifying Data Based on Conditions

• Change Values: df.loc[df['Type 1'] == 'Fire', 'Type 1'] = 'Flamer'
• Set Multiple Columns: df.loc[df['Total'] > 500, ['Generation', 'Legendary']] = ['Test 1', 'Test 2']

Group By and Aggregate Functions

• Group By: df.groupby(['Type 1']).mean().sort_values('Defense', ascending=False)
• Sum and Count: df.groupby(['Type 1']).sum(), df.groupby(['Type 1']).count()

Handling Large Datasets

• Read in Chunks: for chunk in pd.read_csv('modified.csv', chunksize=10000)

Conclusion

• Subscribe: For more tutorials and advanced topics
• Feedback: Leave comments for clarity and additional features