Data Science with Python

Python basics, statistics, machine learning, NLP and deep learning, with code you run in the browser.

Comes after: SQL for Data Analysis

1. Python Basics

2. Python Variables and Data TypesStore values in variables and meet the four basic types: int, float, str and bool.
3. Python Lists and LoopsKeep many values in a list, then process them with for loops and list comprehensions.
4. Python DictionariesStore data as key-value pairs: the structure behind JSON, records and model settings.
5. Python FunctionsPackage code into reusable functions with parameters, default values and return values.
6. NumPy ArraysDo fast maths on whole arrays at once with NumPy, the foundation of the Python data stack.
7. pandas DataFramesLoad, filter, group and summarise tabular data with pandas.
8. Data Visualization with MatplotlibTurn a DataFrame into line, bar, scatter and histogram charts, and learn what each is for.

9. Statistics

10. Descriptive StatisticsSummarise a dataset with measures of centre and spread, find outliers and read a five-number summary.
11. Probability and DistributionsLearn the rules of probability and the two distributions you will meet most: the normal and the binomial.
12. Sampling and Confidence IntervalsSee why averages of samples behave predictably and how to state an estimate together with its uncertainty.
13. Hypothesis TestingDecide whether a difference in your data is real or could be chance, using p-values, t-tests and chi-square tests.
14. Correlation and Simple Linear RegressionMeasure how strongly two variables move together and fit a straight line to predict one from the other.

15. Machine Learning

16. The Machine Learning WorkflowTrain your first model with scikit-learn and learn the steps every project follows, from features to test score.
17. Linear RegressionPredict a number from several features, read the coefficients and measure the error with MAE, RMSE and R squared.
18. Classification with Logistic RegressionPredict a category and its probability, scale the features and choose a decision threshold.
19. Evaluating Models: Precision, Recall and Cross-ValidationGo beyond accuracy with the confusion matrix, precision, recall, F1 and ROC AUC, and get a reliable score with cross-validation.
20. Decision Trees and Random ForestsLearn the models that power most tabular machine learning, see a tree overfit and fix it with a forest.
21. Feature Engineering and PipelinesHandle missing values, encode categories and scale numbers inside a pipeline that prevents data leakage.
22. Overfitting, Underfitting and Hyperparameter TuningDiagnose a model that is too simple or too complex and find better settings with grid search.
23. Unsupervised Learning: Clustering and PCAFind groups in unlabelled data with k-means and reduce many features to a few with principal component analysis.

24. Natural Language Processing

25. Text PreprocessingTurn raw text into clean tokens a program can count: lowercase, tokenise, remove stop words and normalise.
26. Bag of Words and TF-IDFConvert documents into numeric vectors a model can use, and measure how similar two texts are.
27. Text ClassificationTrain a model that labels text, such as positive or negative reviews, and see which words drive its decisions.
28. Word EmbeddingsRepresent words as vectors that capture meaning, so that similar words are close together.
29. Transformers for NLPUnderstand attention, the mechanism behind modern language models, and use a pretrained model in a few lines.

30. Deep Learning

31. Neural Network BasicsBuild a neuron, an activation function and a layer in NumPy, and see how data flows through a network.
32. Training Neural NetworksLearn how a network improves: loss functions, gradient descent and backpropagation, with a network trained from scratch.
33. PyTorch BasicsWrite a real training loop with tensors, automatic gradients and a model class, the way deep learning is done in practice.
34. Convolutional Neural NetworksSee how networks recognise images: filters that detect edges, pooling that shrinks the picture, and layers that build up to objects.
35. Sequence Models and the Road to TransformersFollow the path from recurrent networks to attention, and see how deep learning leads into generative AI.