Python Data Analysis From Scratch: NumPy, Pandas, Seaborn

Learn Data Analysis From Scratch Using Python takes you from setting up Jupyter Notebook to running a full exploratory data analysis (EDA) with NumPy, Pandas, Matplotlib and Seaborn. It is for beginners who want to become data analysts, students starting data science, and professionals who work in spreadsheets today and want to analyse data with Python.

The course starts with a Python refresher, then covers NumPy, Pandas, core statistics, Matplotlib and Seaborn, more than twenty plot types grouped by purpose, and a seven-step EDA workflow. It is a long, complete course of 80 videos, about 11.2 hours in total, free on YouTube and listed on the Python data analysis course page.

Who this course is for

• Aspiring data analysts who want one structured path instead of scattered tutorials.
• Students in data science, statistics or engineering who need Python skills for assignments and projects.
• Excel users and business professionals ready to handle bigger datasets and repeatable analysis.
• Developers who know some Python but have not used NumPy, Pandas or Seaborn.

The catalog marks it Beginner. A short prerequisites video opens the course, and a Python refresher covers the language features you need.

What you will learn

1. Setup and a Python refresher. The IPython interpreter, Jupyter Notebook, basic data types, lists, dictionaries, sets, tuples, functions, and classes and objects. Every later lesson runs in Jupyter, so getting comfortable there first makes the rest smoother.

2. NumPy, the numerical core. What NumPy is and why to use it, ways to create arrays, how arrays work internally, data types and casting, slicing and indexing, array operations, broadcasting and vectorisation. Vectorised thinking is what makes Python analysis fast, and Pandas is built on it.

3. Pandas for real data work. Creating DataFrames, DataFrame basics, GroupBy, handling missing data, aggregation, window functions, transform, filter, join, merge and concatenate, the apply method, reshaping and frequency distributions. This is the toolkit you will use every day as an analyst.

4. Data and statistics concepts. What data is, datasets, types of variables, types of statistics, the normal distribution, mean, median and mode, measures of variability, the Z score, and modality, skewness and kurtosis. These ideas tell you what your numbers actually mean before you plot them.

5. Visualisation with Matplotlib and Seaborn. The Matplotlib architecture and plot parameters, a Seaborn overview, then plot types organised by purpose: correlation (scatter, best-fit line, counts, marginal boxplot, correlogram heatmap, pairwise), deviation (diverging bars and dots), ranking (ordered bars, dot plot), distribution (histograms, density, box plot), composition (pie, tree map, bar) and change over time (time series and decomposition). Choosing a chart by the question you are asking is one of the most valuable habits in the course.

6. The EDA workflow. The big picture, what EDA is, and seven steps: variable identification, univariate analysis, bivariate analysis, missing value treatment, outlier detection and treatment, variable transformation and variable creation.

How the course is organised

There are 80 videos and about 11.2 hours in total, mostly between 3 and 15 minutes each. The order runs:

• Videos 1–11: introduction, prerequisites, IPython, Jupyter and the Python refresher.
• Videos 12–21: NumPy (Part II).
• Videos 22–35: Pandas.
• Videos 36–45: data analysis and descriptive statistics concepts.
• Videos 46–69: Matplotlib, Seaborn and plot types by purpose.
• Videos 70–80: the EDA workflow and a closing “What Next?”.

How to study it step by step

1. Install Jupyter with the setup videos and keep one notebook per section.
2. Skim the Python refresher if you already know Python, but do not skip classes and objects if they are new to you.
3. Type every NumPy and Pandas example yourself. After the Pandas block, load a small dataset of your own and repeat GroupBy, missing data handling and merging on it.
4. Treat the statistics block as theory to apply. Calculate the mean, median, standard deviation and Z scores on your dataset.
5. Learn plots by purpose. For each plot type, write down the question it answers, then make it with your own data.
6. Run the seven EDA steps from start to finish on one dataset. That notebook becomes a portfolio piece you can show employers.

This course currently has no Skills Pragati hands-on lab guide, practice simulation or mock exam.

Frequently asked questions

Is the Python data analysis course free?

Yes. All 80 videos are free on YouTube. Follow them in order from the course page.

How long does it take?

About 11.2 hours of video. With practice in Jupyter, most learners need six to eight weeks at a few hours per week.

Do I need to know Python already?

No. The course is rated Beginner and includes a Python refresher covering data types, collections, functions and classes. Some exposure to any programming helps, but it is not required.

What should I study next?

If you want to move towards machine learning, Machine Learning Math Simplified explains the linear algebra and calculus behind ML models, including the PCA and tensor operations that build on NumPy.