🐼 PANDAS FULL COURSE 2025
Complete Educational Guide · Data Analysis with Python · Digital E-Filing Coach – Amanuddin Education
🔵 1. INTRODUCTION TO PANDAS
What is Pandas?
Pandas is like a super-powered Excel inside Python. It lets you load, clean, analyse, and transform data using simple Python commands — no complicated coding needed.
Key Points
- Open Source Python Library — Free to use. Just install it with one command: pip install pandas. No payment needed.
- Built on NumPy — Pandas is built on top of NumPy (another Python library), making it very fast when working with large numbers and datasets.
- Panel Data Derivation — The name "Pandas" comes from "Panel Data" — a term used in statistics for multi-dimensional structured data.
- Data Analysis and Manipulation — You can filter rows, add columns, calculate averages, sort data, and much more — all with just 1–2 lines of code.
Simple Example
import pandas as pd # Step 1: Import Pandas
data = {'Name': ['Aman', 'Sara', 'Ravi'],
'Score': [85, 92, 78]}
df = pd.DataFrame(data) # Create a table
print(df)
📤 Output:Name — Score
Aman — 85 | Sara — 92 | Ravi — 78
Aman — 85 | Sara — 92 | Ravi — 78
| Feature | What it Means | Simple Analogy |
|---|---|---|
| Open Source | Free, public code | Like a free public library |
| Built on NumPy | Uses NumPy arrays inside | Building on a strong foundation |
| Panel Data | Multi-dimensional structured data | Big Excel sheet with rows & columns |
| Data Manipulation | Filter, sort, merge, clean data | Editing a spreadsheet with superpowers |
🟢 2. DATA STRUCTURES
2A. Pandas Series — 1D Labeled Array
A Series is like a single column of an Excel sheet — one list of values, each with a label (index).
- 1D Labeled Array — One-dimensional, like one column of a table.
- Holds Any Data Type — Numbers, text, dates, booleans — anything goes in a Series.
- Labeled Indices — Each value has a label: default 0,1,2 or custom labels like 'a','b','c'.
- Creation via Lists or Dictionaries — Create from a Python list or dictionary easily.
s = pd.Series([10, 20, 30], index=['a','b','c'])
print(s) # a→10, b→20, c→30
2B. Pandas DataFrame — 2D Labeled Table
A DataFrame is like a full Excel table — many rows and columns together. Each column is actually a Series.
- 2D Labeled Table — Has rows and columns, just like a spreadsheet.
- Collection of Series — Every column in a DataFrame is a Series.
- Rows (Observations) — Each row = one data record (e.g., one student, one transaction).
- Columns (Features) — Each column = one attribute (e.g., Name, Age, Score).
df = pd.DataFrame({
'Name' : ['Aman', 'Sara'],
'Age' : [25, 22],
'Score' : [88, 95]
})
| Structure | Dimensions | Analogy | Creation Method |
|---|---|---|---|
| Series | 1D (one column) | Single Excel column | pd.Series([1,2,3]) |
| DataFrame | 2D (rows + columns) | Full Excel sheet | pd.DataFrame({...}) |
🟠 3. DATA ACCESS AND INDEXING
3A. Indexers — Selecting Data
Indexers are used to pick specific rows and columns — like clicking specific cells in Excel.
▸ loc — Label-based Selection
Use loc when you know the row/column name (label).
df.loc[0, 'Name'] # Row 0, column 'Name' → 'Aman'
df.loc[1:2, ['Name','Score']] # Multiple rows and columns
▸ iloc — Integer Position Selection
Use iloc when you know the position number (0, 1, 2...).
df.iloc[0, 1] # Row 0, Column 1 (integer) → 25
df.iloc[0:2, 0:2] # First 2 rows, first 2 columns
▸ at / iat — Single Cell Access (Fastest)
df.at[0, 'Name'] # By label → 'Aman'
df.iat[0, 0] # By position → 'Aman'
3B. Methods
▸ Head and Tail
df.head(3) # First 3 rows
df.tail(3) # Last 3 rows
▸ Boolean Indexing — Filter by Condition
df[df['Score'] > 80] # Only rows where Score is greater than 80
▸ Select Dtypes
df.select_dtypes(include=['int64']) # Only integer columns
| Method | Used For | Example |
|---|---|---|
| loc | Label-based row/col access | df.loc[0,'Name'] |
| iloc | Integer position access | df.iloc[0,1] |
| at | Single label-based cell | df.at[0,'Age'] |
| iat | Single integer-based cell | df.iat[0,1] |
| head(n) | First n rows preview | df.head(5) |
| tail(n) | Last n rows preview | df.tail(5) |
| Boolean | Filter rows by condition | df[df['A']>10] |
| select_dtypes | Filter columns by data type | df.select_dtypes('float') |
🟣 4. KEY OPERATIONS
4A. Arithmetic Operations
- Broadcasting — Apply one value to all rows automatically. Example: add 5 to every student's score.
- Inter-column Operations — Do math between two columns. Example: Total = Maths + Science.
- add, sub, mul, div Methods — Named methods that handle missing values (NaN) more gracefully than + - * /.
df['Score'] + 5 # Broadcasting: adds 5 to every row
df['Total'] = df['Math'] + df['Science'] # Inter-column
df['Score'].add(10, fill_value=0) # Named method
4B. Statistical Operations
- mean, median, mode — Average, middle value, and most repeated value of a column.
- sum, min, max — Total, lowest, and highest values in a column.
- idxMin, idxMax — Returns the row index (label) of the minimum or maximum value.
- describe() — Gives a complete statistical summary in one shot.
df['Score'].mean() # Average score
df['Score'].idxmax() # Which row has the highest score?
df.describe() # Full statistical summary
4C. Data Manipulation
- Sort Values and Index — Arrange rows by a column's value or by the row index.
- AsType Conversion — Convert a column's data type (e.g., text → number).
- Value Counts — Count how many times each unique value appears in a column.
- Unique Values — See only the distinct (non-repeating) values in a column.
- Inplace Parameter — inplace=True saves the change directly to the DataFrame, no need to reassign.
df.sort_values('Score', ascending=False) # Descending sort
df['Age'] = df['Age'].astype(int) # Convert to integer
df['City'].value_counts() # Count each city
df['City'].unique() # Unique cities only
df.drop_duplicates(inplace=True) # Remove duplicates directly
| Operation | Method | Simple Meaning |
|---|---|---|
| Mean | .mean() | Average of all values |
| Median | .median() | Middle value when sorted |
| Mode | .mode() | Most frequently occurring value |
| Sum / Min / Max | .sum() .min() .max() | Total / Smallest / Largest |
| idxmin / idxmax | .idxmin() .idxmax() | Row index of min/max value |
| Describe | .describe() | Full statistics summary report |
| Sort Values | .sort_values() | Arrange rows by column value |
| AsType | .astype() | Change column data type |
| Value Counts | .value_counts() | Count each unique value |
| Unique | .unique() | List of all distinct values |
| Inplace | inplace=True | Save change to same object |
🟡 5. ADVANCED FUNCTIONS
5A. Group By
groupby() splits your data into groups and lets you analyse each group separately — exactly like Pivot Table in Excel.
df.groupby('City')['Score'].mean()
# Average score for each city separately
📍 Real Example:If data has students from Delhi and Mumbai, groupby('City').mean() gives the average score of Delhi students AND Mumbai students — separately in one line.
5B. Aggregate (agg)
agg() lets you apply multiple statistics at once on grouped data — saves you running many commands separately.
df.groupby('City')['Score'].agg(['mean','max','min','count'])
# Shows mean, max, min, count for each city — all at once
5C. Apply and Lambda
apply() runs your own custom function on every row or column. Lambda is a one-line shortcut function — no need to write a full def block.
df['Grade'] = df['Score'].apply(lambda x: 'Pass' if x>=50 else 'Fail')
# Adds a Grade column: 'Pass' if score ≥ 50, otherwise 'Fail'
5D. ApplyMap
applymap() applies a function to every single cell in the entire DataFrame — useful for formatting or transforming all values.
df.applymap(lambda x: round(x, 2)) # Round every cell to 2 decimals
| Function | Works On | Purpose |
|---|---|---|
| groupby() | DataFrame | Split data into groups for separate analysis |
| agg() | GroupBy result | Multiple statistics applied at once |
| apply() | Column or Row | Apply any custom function |
| lambda | Inline in apply() | One-line function without def |
| applymap() | Entire DataFrame | Function applied to every individual cell |
🔷 6. DATA LOADING
6A. read_csv — Load a CSV File
Loads data from a CSV file (Comma Separated Values — a plain text file where data is separated by commas, like when you export from Excel).
df = pd.read_csv('students.csv')
print(df.head()) # Preview first 5 rows
6B. read_excel — Load an Excel File
Loads data from an Excel file (.xlsx or .xls). You can also choose which specific sheet to load.
df = pd.read_excel('marks.xlsx', sheet_name='Sheet1')
6C. index_col — Set a Column as Row Index
Tells Pandas which column should become the row label (index) instead of the default 0, 1, 2, 3...
df = pd.read_csv('data.csv', index_col='RollNo')
# Now RollNo column becomes the row index label
📍 Simple Example:If your CSV has a "StudentID" column, using index_col='StudentID' makes each student's ID the row label, so you can easily search by df.loc['S001'].
| Function | File Type | Key Parameters | Quick Example |
|---|---|---|---|
| read_csv() | CSV (.csv) | sep, header, dtype, nrows | pd.read_csv('file.csv') |
| read_excel() | Excel (.xlsx/.xls) | sheet_name, usecols, skiprows | pd.read_excel('file.xlsx') |
| index_col | Both (CSV & Excel) | Column name or column number | index_col='RollNo' |
📊 PANDAS LEARNING FLOWCHART
Step-by-step learning path for Pandas — from installation to professional data analysis.
🐍 STEP 1 — Install Python & Pandas
pip install pandas numpy
pip install pandas numpy
📦 STEP 2 — Import Pandas
import pandas as pd
import pandas as pd
📂 STEP 3 — Load Your Data
read_csv() / read_excel() / index_col
read_csv() / read_excel() / index_col
🗂️ STEP 4 — Understand Data Structures
Series (1D) & DataFrame (2D)
Series (1D) & DataFrame (2D)
🔍 STEP 5 — Access & Filter Data
loc / iloc / at / iat / Boolean Indexing
loc / iloc / at / iat / Boolean Indexing
📐 STEP 6 — Arithmetic Operations
Broadcasting / add / sub / mul / div
Broadcasting / add / sub / mul / div
📊 STEP 7 — Statistical Analysis
mean / median / mode / describe / idxmax
mean / median / mode / describe / idxmax
🔧 STEP 8 — Data Manipulation
sort / astype / value_counts / unique / inplace
sort / astype / value_counts / unique / inplace
⚙️ STEP 9 — Advanced Functions
groupby / agg / apply / lambda / applymap
groupby / agg / apply / lambda / applymap
📈 STEP 10 — Visualise & Export
df.plot() / to_csv() / to_excel()
df.plot() / to_csv() / to_excel()
✅ DATA ANALYSIS COMPLETE!
Insights ready for decisions 🎯
Insights ready for decisions 🎯
🧠 PANDAS MIND MAP
Complete visual overview of all Pandas topics from the mind map.
🗺️ PANDAS LEARNING ROADMAP 2025
Structured 6-Phase Roadmap — follow week by week to master Pandas from zero to professional level.
🔵 PHASE 1 — Foundation (Week 1)
- Install Python 3.x and Pandas using pip
- Understand what Pandas is and why data analysts use it
- Learn NumPy basics — the foundation beneath Pandas
- Create your first Series and DataFrame from lists and dictionaries
- Understand index, columns, values, shape, and dtypes
🟢 PHASE 2 — Data Loading (Week 2)
- Load CSV files using read_csv() with all key parameters
- Load Excel files using read_excel() — choose sheets and columns
- Use index_col, header, usecols, dtype, nrows parameters
- Explore loaded data: head(), tail(), info(), shape, describe()
- Handle missing values: isnull(), dropna(), fillna()
🟠 PHASE 3 — Data Access & Indexing (Week 3)
- Master loc (label-based) and iloc (integer-based) selection
- Use at and iat for fast single-cell access
- Filter rows using Boolean Indexing with conditions
- Select specific column types with select_dtypes()
- Understand head(), tail(), and slicing techniques
🟣 PHASE 4 — Operations & Statistics (Week 4)
- Arithmetic: Broadcasting and inter-column math operations
- Use add(), sub(), mul(), div() named methods safely
- Calculate mean, median, mode, sum, min, max statistics
- Use idxmin(), idxmax(), describe() for quick summaries
- Sort with sort_values() and sort_index()
- Convert data types with astype() properly
- Analyse with value_counts() and unique()
🔴 PHASE 5 — Advanced Functions (Week 5)
- Master groupby() for grouped data analysis (like Pivot Table)
- Use agg() for multiple statistics on groups at once
- Write custom logic with apply() and lambda functions
- Apply cell-level transformations with applymap()
- Merge, join, and concatenate multiple DataFrames
- Create Pivot Tables and cross-tabulations
🔷 PHASE 6 — Real Projects & Export (Week 6)
- Work on real-world datasets — Sales, Marks, Finance, GST data
- Visualise data with df.plot() and matplotlib integration
- Export clean results to CSV with to_csv() and Excel with to_excel()
- Build an end-to-end data analysis mini-project from scratch
- Present and share analysis reports professionally
⚠️ Educational Disclaimer: This resource is for educational purposes only and does not constitute legal or professional advice.
All code examples are for learning purposes only. Refer to official Pandas documentation at pandas.pydata.org for authoritative guidance.
© 2025 Digital E-Filing Coach – Amanuddin Education. All Rights Reserved.
© 2025 Digital E-Filing Coach – Amanuddin Education. All Rights Reserved.