In this machine learning course you learn to build a verifiable predictive model from tabular data with Python and scikit-learn, and it is aimed at participants with Python experience. You prepare data in pipelines, train classification and regression models, evaluate them with cross-validation and suitable metrics and recognise data leakage and overfitting. The course is hands-on with notebooks and explicitly explains where models fail. Deep learning and large language models are not in scope.
What you’ll take away.
- Clean data with pandas and preprocess it in scikit-learn pipelines
- Train and compare classification and regression models
- Evaluate models with cross-validation and suitable metrics such as F1 or MAE
- Recognise and fix data leakage, overfitting and class imbalance
- Document a model traceably and save it for deployment
Your training in detail
- Who is it for?
- Analysts, developers and specialists with Python skills who want to build predictive models from company data.
- Prerequisites
- Confident Python including functions and pandas, plus basic statistics such as mean, spread and correlation.
- Format & planning
- Suggested format: 3 days (about 7 hours each) live online or on site by arrangement. Scope depends on prior knowledge.
Agenda: from understanding to application
Our independently developed agenda proposal. Focus areas, lab environment and scope are agreed before booking.
Problem and data: classification versus regression, defining the target, setting a baseline model and success measure.
Data preparation with pandas and scikit-learn: missing values, encoding categorical features, scaling and ColumnTransformer.
Comparing models: linear and logistic regression, decision trees, random forests and gradient boosting.
Evaluation: train-test split, cross-validation, confusion matrix, precision, recall, F1 and error measures for regression.
Pitfalls: data leakage from preprocessing before the split, overfitting, random seeds and class imbalance.
Tuning and deployment: hyperparameter search with GridSearchCV, feature importance, saving the model and documenting limits.
Suggested exercise: Predict one feature in a public sample dataset, compare two models by cross-validation and write a note describing where the model is unreliable.
Exams, licences & credentials
No specific exam is associated with this course. Attending alone awards no vendor certification. Exam fees are only included if confirmed in the offer.
Vendor references & review date
Sources checked on 2026-11-03. These sources describe the technology or external exam requirements. They do not imply a partnership with NextNowa.
Your learning content
These objectives describe the training topic. The specific syllabus, duration, dates, trainer and price are agreed before booking.
Frequently asked questions
What prior knowledge do I need for the machine learning course?
You should be able to write Python confidently and load and filter tables with pandas. Some statistics such as mean, spread and correlation helps. Advanced mathematics is not required. If you are only just learning Python, take the practical Python course first and this one afterwards.
Does the course also cover deep learning or large language models?
No. The course focuses on classical machine learning with tabular data in scikit-learn, which is often sufficient in companies. Neural networks and language models are separate topics with different tools. If you need them, contact us and we will plan a suitable format.
How do I know whether my model really works?
Test it only on data the model has never seen and compare it with a simple baseline. Cross-validation shows how much results vary. A suspiciously good result is often a data leak. You practise exactly this review routine in the course on your own examples.
Does Machine learning with Python include a certificate?
No specific exam is associated with this course. Attending alone awards no vendor certification. Exam fees are only included if confirmed in the offer.