Exploratory Data Analysis Course with Python and pandas
Exploratory Data Analysis is an intermediate-level course of 5 lessons (about 3 h 1 min), with 15 exercises and 15 quiz questions. Prerequisites: Data Cleaning and Preparation, Descriptive Statistics. Exercises run in the browser, with nothing to install.
- Data analysis
- Intermediate
- 5 lessons
- 3 h 1 min
- 15 exercises
- Exercises run in the browser
The course in brief
This course gives you a complete exploratory analysis method with pandas and matplotlib. You will learn to profile a dataset, handle missing values and duplicates, describe distributions and categories, cross variables without falling into Simpson's paradox, then check consistency, explore time and write an actionable summary.
What you will be able to do
- Profile a dataset and fix data types with pandas 3
- Measure, explain and handle missing values and duplicates
- Describe the distribution of a numeric or categorical variable
- Cross variables with groupby, pivot_table, crosstab and correlations
- Detect business anomalies and explore a change over time
- Write a prioritized, actionable EDA summary
Syllabus: 5 lessons
- Profile a DataFrame (dimensions, types, preview, statistical summary)
- Fix wrongly imported column types (numbers, dates)
- Formulate analysis questions before computing
- 02
Missing values and duplicates
- Measure and locate missing values
- Choose a handling strategy (removal, imputation, indicator)
- Detect and handle exact and key-based duplicates
- 03
Univariate analysis: distributions and categories
- Describe the distribution of a numeric variable (center, spread, skewness, shape)
- Summarize a categorical variable and group rare categories
- Plot a histogram and a bar chart with matplotlib
- 04
Bivariate analysis: crossing variables
- Compare a numeric variable across groups with groupby and pivot_table
- Measure the links between numeric variables with a correlation matrix
- Recognize Simpson's paradox and structure effects
- 05
Anomalies, change over time and the EDA summary
- Automate business consistency checks
- Explore a change over time (resample, moving average, seasonality)
- Write a prioritized, actionable EDA summary
Final project
Complete EDA of a multichannel sales history
Generate in Python (fixed seed) an 18-month order history for a retailer: date, region, channel, category, quantity, unit price, amount, customer. Deliberately introduce realistic defects (amounts as text with a comma, invalid dates, key-based duplicates, missing values concentrated in one region, a few inconsistencies) and a structure effect between channels. Carry out the complete EDA and write a summary.
Deliverables
- A commented notebook or script, executable end to end
- A quality report (fixed types, missing values, duplicates, consistency checks) with the treatments applied
- The key univariate and bivariate analyses with at least 4 readable charts
- A time analysis (monthly, day of the week, moving average)
- A one-page summary: quality, 3 to 5 quantified findings, hypotheses, next steps
Assessment criteria
- The grain and types are correctly identified and handled
- Quality treatments are justified and documented, with no silent correction
- Group comparisons display the counts and account for structure effects
- Charts are suited to the variable types and readable
- The summary is prioritized, quantified and distinguishes findings from hypotheses
Frequently asked questions
What is exploratory data analysis?
It is the step where you get to know a dataset: structure, quality, distributions and relationships between variables, in order to formulate findings and hypotheses before any modeling or decision.
What are the prerequisites for this course?
Knowing how to manipulate and clean data with pandas and knowing the basics of descriptive statistics (mean, median, standard deviation, quartiles).
Is the course up to date with pandas 3?
Yes: it uses pandas 3.1 and takes into account the default 'str' string type and the copy-on-write introduced in pandas 3.
What will I produce at the end?
A complete exploratory analysis of a sales dataset, with quality checks, charts and a written summary aimed at a decision maker.
This course, told for your job
Say what you want to be able to do: “Exploratory Data Analysis” takes its place in your path, with examples from your field. The concepts stay those of the course.