Skilling projects

The 24 skilling sessions, each applying a course technique to a public dataset, with the function to write and the topic to study.

The skilling sessions apply what you learned on PlacementPredict to other public datasets. Do them on your own computer, or in a free online notebook, because the datasets are loaded from the internet. Your instructor will confirm which dataset to use where it says "set by your instructor".

Most of these datasets can be fetched by name with scikit-learn's fetch_openml. Check the page of each one for its terms of use.

Session	Project	Dataset	Function to write	Study
1	EDA and lifecycle mapping	Titanic	A 4-page PDF EDA report	Topic 4
2	Preprocessing pipeline	Titanic	`validate_pipeline(pipeline, X_train, X_val)`	Topic 8
3	Feature engineering and EDA	Adult census	`adult_eda_report(df)`	Topics 4, 8
4	Linear regression and regularisation	Set by your instructor	`regression_benchmark(X_train, y_train, X_test, y_test, alphas)`	Topics 5, 7
5	Full logistic regression pipeline	Set by your instructor	`logistic_pipeline_report(...)`	Topic 6
6	Multinomial logistic regression	Glucose bands (diabetes data)	`multinomial_analysis(...)`	Topic 6
7	Regression, scaling and encoding	Auto MPG	`regression_pipeline(df, target_col="mpg")`	Topics 5, 8
8	Decision tree classification	Titanic	`dt_full_analysis(...)`	Topic 9
9	Random forest ensemble	Set by your instructor	`ensemble_comparison(...)`	Topic 10
10	XGBoost, LightGBM and SHAP	Set by your instructor	`shap_feature_summary(model, X_val, feature_names, top_n=5)`	Topics 11, 12
11	Trees and forests	Wine quality	`model_comparison(...)`	Topics 9, 10
12	Boosting and SHAP	Set by your instructor	`boosting_pipeline_report(...)`	Topics 11, 12
13	Clustering	Customer segments	`cluster_report(df, feature_cols, k)`	Topics 13, 14
14	PCA, t-SNE and UMAP	Customer segments	`compare_dr_methods(X_scaled, labels, perplexity=30, n_neighbors=10)`	Topic 15
15	Isolation Forest	Set by your instructor	`anomaly_benchmark(X, y_true, contamination_values)`	Topic 16
16	Clustering and PCA profiling	Socio-economic data	`profile(df, feature_cols, k, region_col=None)`	Topics 13, 15
17	DBSCAN and Isolation Forest	Network intrusion data	`intrusion_detection_report(...)`	Topics 14, 16
18	CV, metrics and calibration	Set by your instructor	`calibration_comparison(models_dict, ...)`	Topics 17, 20
19	Tuning and significance	Set by your instructor	`tuning_summary(...)`	Topics 19, 20
20	Threshold tuning and calibration	Set by your instructor	`threshold_analysis(model, X_val, y_val, recall_target=0.95)`	Topics 18, 20
21	Search, validation and learning curves	Banknote authentication	`search_comparison(estimator, param_grid, X_train, y_train, n_iter=30)`	Topic 19
22	Modular pipeline and FastAPI	Set by your instructor	`api_health_check(base_url, test_payloads)`	Topics 21, 23
23	MLflow, DVC and drift	Set by your instructor	`psi_report(reference_df, production_df, num_cols, threshold=0.2)`	Topic 24
24	Batch scoring, PSI and retraining	Wine quality	`retrain_pipeline(reference_csv, production_csv, model_path, psi_threshold=0.2)`	Topics 24, 25