Part of the Machine Learning JourneyNewJourneyUnsupervised Learning in Python
Find the groups, the shape and the odd ones out in data that has no labels. Cluster, shrink and scan real datasets with scikit-learn, and see every result on a chart.
51 lessons51 challenges210 quiz questions
- Beginner friendly
AI-assisted coding help
Hands-on interactive lessons
Quizzes to test your knowledge
Free certificate of completion
Syllabus
This section is part of the Machine Learning Journey. The full syllabus has more sections: click any preview below to view it on the Journey page.
- Section 1Data and Plots60 lessons
- Section 2Statistics for ML64 lessons
- Section 3Supervised Learning58 lessons
Section 4
Unsupervised Learning
Start sectionStartFind structure in data without labels: k-means, hierarchical and density clustering, PCA to see and compress many features, and anomaly detection, ending with a digits exploration project.Finding Groups
5 lessons520Choosing k
5 lessons520Reading Clusters
5 lessons520Hierarchical Clustering
4 lessons415Density Clustering
4 lessons415Fewer Dimensions
5 lessons520PCA in Practice
5 lessons520Anomaly Detection
5 lessons520Clusters at Work
4 lessons415Project: Digit Explorer
5 lessons525Final Challenges
4 lessons420
Learn by Doing
Write real code, query databases, build websites, and master AI prompts. Our interactive lessons cover every skill modern developers need.
Build Your Streak
Stay consistent and watch your progress grow! Track your daily coding habit, protect your streak with freeze days, and earn rewards for showing up every day.
12 days streak
Return tomorrow to keep your streak!
January 2026
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
21
22
23
24
25
26
27
28
29
30
Double or Nothing
Day 5 of 7
Streak Freeze
2 left
Code Anywhere
Take your coding journey on the go! No setup, no downloads: just open and start coding. Available on iOS, Android and Web with 4.9 star ratings.
You're Not Alone
Compete on global leaderboards, invite friends to earn rewards, and celebrate each other's wins. Coding is better with friends!
1
2
3
Promotion zoneEvery way to learn
Read, listen, test yourself, ask the AI, or look up anything you've already covered. Every lesson meets you where you are.
A variable is a named container that stores a value you can reference later in your program.
In Python, you create one by writing the name, an equals sign, then the value you want to store.
The value can change over time. Reassigning the name simply points it to a new value.
Earn a Certificate
Earn certificates for every course you complete. Add them to your LinkedIn profile and resume to showcase your coding expertise to employers.
Certificate of CompletionPython Fundamentals
Add to LinkedInWhy learn unsupervised learning with Coddy
- Clustering you can see. Watch k-means move its centers, read a dendrogram's merge heights, and see DBSCAN find moon-shaped groups that k-means cuts in half. Every lesson has a chart, and yours appears in the Plots tab as you run.
- The core methods: k-means, choosing k with the elbow and silhouette scores, hierarchical clustering, DBSCAN, PCA and explained variance, isolation forests and local outlier factor, all with scikit-learn.
- Results you can explain. You profile each cluster in real units, compare clusters with known labels, and name the segments a business would act on, instead of stopping at a list of cluster numbers.
- 64 dimensions made visible: handwritten digits projected onto a flat map, PCA components drawn as images, and pictures rebuilt from a few numbers.
Frequently asked questions about unsupervised learning
What is unsupervised learning?
Unsupervised learning finds structure in data that has no labels. Nobody tells the model the right answer; it looks for groups of similar rows (clustering), for a few directions that keep most of the information (dimensionality reduction), or for rows unlike the rest (anomaly detection). It is how you explore a new dataset, segment customers or spot fraud before anyone has labelled it.
What are examples of unsupervised learning?
Grouping shoppers by what they buy, reducing an image's colours to the eight that matter, mapping 64-pixel digit images onto a flat chart, finding unusual server readings, and splitting cell measurements into two kinds without being told which is which. The course builds each of these.
Which algorithms does the course cover?
k-means with the elbow method and silhouette scores, agglomerative (hierarchical) clustering with dendrograms, DBSCAN, PCA, isolation forests and local outlier factor. You also learn why scaling comes first and when each method fails, so you can pick the right one.
How do I choose the number of clusters?
There is rarely one right answer, so the course teaches three ways to look: the elbow of the inertia curve, silhouette scores, and a dendrogram's merge heights. Then you check the clusters in real units and ask whether a person could name them. A k that gives groups nobody can describe is the wrong k.
Should I learn supervised learning first?
How long does the unsupervised learning section take?
51 lessons, about 8 to 9 hours at your own pace: a few weeks at half an hour a day, including the digit explorer project and four final challenges.
Can I learn unsupervised learning for free?
Yes. The lessons, the charts, the AI hints and the section certificate are free. A free account works through a few lessons a day, and Pro removes the daily limit.
