About Projects Courses Mind Map Notes Roadmap Blog Services

Loading…

Newsletter

New tutorials, notes & project breakdowns

Practical ML, GenAI and full-stack lessons, straight to your inbox. No spam, unsubscribe anytime.

Munna Kumar, also known as Code With Munna (CodeWithMunna / CodeWithMunnaX). Data science, ML and GenAI engineering, built for production and taught for free on YouTube.

Explore

About
Projects
Skills
Blog
Tech Setup

Learn

Free Courses
Python Notes
NumPy Notes
Pandas Notes
SQL Notes
YouTube Channel

Work With Me

Services
How I Work
Book a Free Call
Contact

© 2026 Munna Kumar · Made in India

Loading notes

NotesPandas

🐼 Free notes

Pandas Notes

Pandas from scratch to real projects — Series, DataFrames, loc/iloc, cleaning, GroupBy, pivot, merge and dates, with a live Python playground on every topic. Written by Munna Kumar (Code With Munna) — 83 topics in order, each with examples, key points, and 33 with a video lesson.

Start with What is Pandas?

Getting Started

1What is Pandas?What pandas is, the kind of data it is built for (labelled, tabular, mixed-type), how it relates to NumPy, and why every data analyst uses it.
2Installing & ImportingHow to install pandas with pip or conda, the standard `import pandas as pd` convention, checking the version, and setting display options.
3Series vs DataFrameA first look at pandas' two core structures — the 1D Series and the 2D DataFrame — how they relate, and how to tell which one you have.
4What's New in pandas 3The pandas 3 changes you will notice when reading older tutorials: the str dtype, Copy-on-Write, applymap → map, inplace guidance and new frequency aliases.

Series

5Creating a SeriesEvery way to build a Series — from a list, tuple, range, NumPy array, dictionary or a single scalar — plus custom index labels, a name and a forced dtype.
6Index & LabelsAccess Series values by label or by position, slice with .loc and .iloc, filter with conditions, and get the index and values out.
7Series AttributesRead the facts about a Series — shape, size, dtype, name, ndim, is_unique, hasnans, empty — and summarise it with info().
8Vectorized Operations & AlignmentDo maths on whole Series at once, and understand index alignment — pandas matches values by label, producing NaN where labels don't match.
9Exploring & Sorting a SeriesPeek at data with head/tail/sample, count categories with value_counts, find distinct values with unique/nunique, and order with sort_values/sort_index.
10Stats, Cumulative & RoundingFind who scored the most with idxmax/idxmin, summarise with describe and agg, track running totals with cumsum/cumprod/cummax/cummin, and tidy values with clip and round.

DataFrames

11Creating a DataFrameBuild DataFrames from a dict of lists, a list of dicts, a list of lists with column names, a NumPy array, or several Series — and know what each one does with missing data.
12DataFrame Anatomy & AttributesThe parts of a DataFrame — index, columns, values, axes — and the attributes that describe it: shape, size, ndim, dtypes, to_numpy and more.
13Inspecting DataThe first five commands on any new dataset: head, tail, sample, info and describe (including include='all'), plus memory_usage.
14Selecting & Adding ColumnsSelect one column (Series) or many (DataFrame), and create new or calculated columns with assignment, np.where, insert and assign.
15Rename, Drop & ReorderRename columns and index labels, remove columns or rows with drop, del and pop, and reorder columns into the order you want.
16Data Types & astypeThe main pandas dtypes (str, int64, float64, bool, category, datetime64), converting with astype, and safely parsing messy numbers with pd.to_numeric(errors='coerce').

Reading & Writing Data

17Reading CSV Files (read_csv)Load a CSV file into a DataFrame with pd.read_csv — file paths, simulating files with io.StringIO, and the first four parameters: sep, header, names and skiprows.
18read_csv Parameters in DepthControl exactly what read_csv loads: index_col, usecols, dtype, parse_dates, na_values, nrows, skiprows, encoding and thousands.
19Reading JSON, Excel, HTML & SQLLoad data from JSON (read_json, lines=True, json_normalize), Excel workbooks (read_excel, sheet_name, ExcelFile), web tables (read_html) and databases (read_sql).
20Writing Data: to_csv & to_jsonSave a DataFrame as CSV or JSON — why index=False matters, the useful to_csv options, and every to_json orient explained.
21to_dict, to_html & to_excelTurn a DataFrame into Python dictionaries (every to_dict orient), an HTML table for a web page, or an Excel workbook with to_excel.
22First Look at a New DatasetA step-by-step checklist for exploring any freshly loaded dataset: shape, head/tail/sample, columns, info, dtypes, describe, isna().sum(), missing %, nunique and value_counts.

Selecting & Filtering

23Selecting ColumnsPick one column (a Series) or several columns (a DataFrame) with df['col'] and df[['a', 'b']], and learn why attribute access df.col is risky.
24loc: Select by LabelUse .loc[rows, columns] to select rows, columns, slices and single values by their labels — with end-inclusive slicing and boolean masks.
25iloc & loc vs ilocUse .iloc[rows, columns] to select by integer position (end-exclusive, negative indexes allowed), and master the full loc vs iloc comparison.
26at & iat: Fast Single ValuesRead or write exactly one cell with .at (by label) and .iat (by position) — the faster scalar versions of loc and iloc.
27Boolean FilteringFilter rows with conditions: build a boolean mask, combine conditions with & | ~ (and why the parentheses are compulsory), and count matching rows.
28isin, between & str.containsShortcut filters: isin for 'one of these values', between for numeric ranges, and a first look at str.contains / startswith for text matching.
29The query() MethodFilter rows with a readable string expression — df.query("team == 'MI' and runs > 100") — using and/or/not, in, chained comparisons, @variables and backticks.
30set_index & reset_indexTurn a column into the row index with set_index (drop, append, multi-column) and turn the index back into a column with reset_index (drop=True for a clean 0..n-1).
31Sorting: sort_values, sort_index, nlargestOrder rows by one or many columns with sort_values (ascending lists, na_position, ignore_index), by labels with sort_index, and grab top-N rows with nlargest / nsmallest.
32Modifying Values Safely (Copy-on-Write)Change cells, columns and filtered rows correctly with df.loc[mask, 'col'] = value, and understand why chained assignment never works under pandas 3's Copy-on-Write.

Data Cleaning

33Detecting Missing ValuesWhat a missing value is in pandas (NaN, None, NaT, pd.NA), and how to find, count and measure the percentage of missing data with isna(), notna() and info().
34Dropping Missing Data (dropna)How dropna() removes rows or columns with missing values, and how axis, how, thresh and subset control exactly what gets dropped.
35Filling Missing Data (fillna, ffill, interpolate)Replace missing values instead of deleting them: fillna with a constant, a per-column dict, mean/median/mode, plus forward fill, backward fill and interpolation.
36Finding & Removing DuplicatesDetect repeated rows with duplicated(), remove them with drop_duplicates(), and control which copy survives with subset and keep.
37Replacing & Fixing ValuesFix wrong or inconsistent values with replace(), recode categories with map(), tidy messy text and rename columns.
38Fixing Data Types & OutliersConvert columns to the right dtype with astype() and pd.to_numeric(errors='coerce'), then detect and handle outliers with the IQR rule and clip().
39End-to-End Cleaning: Messy IPL DataPut every cleaning tool together on a messy IPL matches table: audit, fill, drop, standardise, fix types, de-duplicate and add date features.

Text & Strings

40The .str AccessorWhy pandas text columns need the .str accessor, how it applies Python string methods to every value at once, and how it handles missing values safely.
41Case & Whitespace CleaningStandardise text with lower, upper, title and capitalize, and remove unwanted spaces with strip, lstrip, rstrip and a regex that collapses repeated spaces.
42Searching Text (contains, startswith, match)Find rows whose text contains, starts with or ends with a pattern, locate and count substrings, and control case, missing values and regex behaviour.
43Replacing & Extracting TextRewrite text with str.replace (plain or regex), pull out pieces with str.extract and capture groups, and slice, measure and pad strings.
44Splitting & Joining TextSplit one text column into many with str.split (expand, n), pick pieces with str.get, and combine text with str.cat, str.join and +.
45Practice: Cleaning Text ColumnsA complete text-cleaning project on car names and IPL player data: normalise case and spaces, fix brand typos, split brand/model, extract numbers and categories.

Dates & Times

46Converting to Datetime (pd.to_datetime)Turn date text into real datetime values with pd.to_datetime: formats, format=, dayfirst, errors='coerce', mixed formats and Unix timestamps.
47The .dt AccessorPull parts out of a datetime column with .dt — year, month, day, day_name, quarter, dayofweek, is_month_end — and format dates as text with strftime.
48Date Arithmetic & TimedeltaSubtract dates to get Timedelta durations, add days or months to dates, find gaps between matches with diff(), and convert durations to days, hours or seconds.
49Generating Dates: date_range & PeriodsCreate regular sequences of dates with pd.date_range (start, end, periods, freq), learn the frequency aliases like D, W, ME, h, and meet Periods for whole months and quarters.
50Time Series: Filtering, Resampling & RollingUse a DatetimeIndex to slice data by date strings, aggregate to weekly or monthly totals with resample(), and smooth data with rolling windows.
51Time Zones (tz_localize & tz_convert)Understand naive vs time-zone-aware datetimes, attach a zone with tz_localize, change it with tz_convert, and handle IST, UTC and UAE match times correctly.

Transforming Data

52apply() on a SeriesRun your own Python function on every value of a Series with Series.apply() — custom ratings, extra arguments, and functions that return several values.
53apply() on a DataFrameDataFrame.apply() with axis=0 (one call per column) vs axis=1 (one call per row), returning a Series to create several columns, and the result_type option.
54map(): Series & DataFrameReplace or transform values with Series.map (dict, function or Series), handle unmatched values and NaN, and use DataFrame.map — the pandas 3 replacement for applymap.
55Lambda Functions in PandasWrite small one-line anonymous functions with lambda and use them inside apply, map, sort_values(key=), loc[] and groupby().agg().
56assign() & Method ChainingCreate new columns with assign() without modifying the original, and build readable step-by-step pipelines with method chaining and pipe().
57Conditional ColumnsCreate columns from conditions the vectorised way: np.where for two outcomes, np.select for many, Series.where/mask to keep or replace values, and loc for in-place updates.
58Iterating: items, iterrows, itertuplesLoop over a DataFrame by column (items), by row as a Series (iterrows) or by row as a namedtuple (itertuples) — what each returns, their traps, and when a loop is acceptable.
59Why Vectorization WinsWhat vectorisation means, a timing comparison of iterrows vs itertuples vs apply vs vectorised code, and how to rewrite common loops the fast way.

GroupBy & Aggregation

60Aggregate FunctionsHow sum, mean, median, min, max, count and std squeeze a whole column (or row) into one number, and how axis=0 vs axis=1 decides the direction.
61GroupBy: Split-Apply-CombineHow groupby() splits rows into groups, applies an aggregate to each group and combines the answers — by one or many columns, with as_index, get_group and loops.
62agg() in DepthRun several aggregates per group with agg(): a list of functions, a dict per column, named aggregation like agg(total=('runs', 'sum')), and custom functions.
63transform() & filter()Put a group-level value (team total, team average, rank) back on every row with transform(), and keep or drop whole groups with filter().
64Counting & DistributionsCount how often each value appears with value_counts() (also as percentages), count distinct values with nunique(), and know size() vs count().
65Binning with pd.cut & pd.qcutTurn continuous numbers (runs, ages, strike rates) into labelled categories with pd.cut (your own edges) and pd.qcut (equal-sized quantile groups) — video 07.
66Correlation & describe() by GroupMeasure how two numeric columns move together with corr() (Pearson/Spearman), read a correlation table, and summarise each group with describe() — notebook 08, video 08.

Pivot & Reshape

67pivot(): Long to WideRotate a long table into a wide one with pivot(index, columns, values) — rows become columns — and understand why it fails on duplicate entries.
68pivot_table(): Reshape + AggregateBuild Excel-style summary tables with pivot_table(values, index, columns, aggfunc, fill_value, margins), and know exactly when to choose it over pivot() — video 08.
69melt(): Wide to LongUnpivot a wide table (one column per season) into tidy long format with melt(id_vars, value_vars, var_name, value_name) — the reverse of pivot — video 09.
70stack(), unstack() & MultiIndexUnderstand MultiIndex (hierarchical) labels and move a level between rows and columns with stack() and unstack() — the engine behind pivot and groupby reshapes.
71crosstab(): Frequency TablesCount how often each combination of two categories occurs with pd.crosstab(), add totals with margins, and turn counts into row/column percentages with normalize.

Merge, Join & Concat

72concat(): Stacking DataFramesGlue DataFrames together top-to-bottom (axis=0) or side-by-side (axis=1) with pd.concat — ignore_index, keys, mismatched columns and join='inner'.
73merge(): Joining on Key ColumnsCombine two tables by matching values in a key column — pd.merge / df.merge with on, left_on/right_on, and the default inner join (SQL JOIN for pandas) — video 10.
74Join Types: inner, left, right, outer, crossWhat how='inner' / 'left' / 'right' / 'outer' / 'cross' keeps and drops, with row-by-row diagrams, a Venn-style comparison table and indicator=True — video 10.
75Merge Options: suffixes, validate & Duplicate KeysHandle overlapping column names with suffixes, understand one-to-many and many-to-many matches (exploding rows), and guard merges with validate=.
76join(): Index-Based JoiningUse DataFrame.join() to attach columns from other DataFrames/Series by matching index labels — how=, lsuffix/rsuffix, on=, and joining several frames at once.
77concat vs merge vs joinPick the right combining tool every time: a comparison table, a decision guide, and the classic mistakes (duplicate keys exploding rows, misaligned axis=1, lost rows).

Real-World Pandas

78The EDA WorkflowA repeatable six-step recipe for exploring any new dataset: ask questions, load, inspect, clean, analyse and summarise — worked end to end on a messy IPL players file.
79Project: IPL Match AnalysisA mini-project inspired by notebook 06: load IPL match results, convert dates, and answer real questions — most wins, win %, toss decisions, wins by season, champions and top venues.
80Project: Car Price AnalysisA mini-project inspired by notebooks 04/05: extract the brand from CarName, fix brand typos, compare average price by brand and fuel, build price bands with pd.cut, and measure correlation with engine size and horsepower.
81Plotting with PandasDraw quick charts straight from a DataFrame or Series with .plot(kind=...) — line, bar, barh, hist, box, kde, area, pie, scatter and hexbin — and know which chart answers which question.
82Performance & Memory TipsMake pandas faster and lighter: measure memory, use the category dtype, downcast numbers, load only needed columns with usecols/dtype, stream big files with chunksize, vectorise, and chain methods.
83Pandas Cheat SheetThe whole Pandas course on one page: tables of the most-used methods for creating, reading, inspecting, selecting, cleaning, transforming, grouping, reshaping, combining and plotting data — with pandas 3 notes.

Looking for another subject?

Python Notes NumPy Notes SQL Notes