Predictive analytics uses historical data, statistics and machine learning to estimate likely future events. It answers questions such as: which customers are about to cancel? How much stock will be needed next month? When is a machine likely to fail? The results are probabilities, not certainties. Their value depends heavily on data quality and on whether the predictions actually feed into decisions and processes.
Predictive modelling is the technical core of predictive analytics: example data with known outcomes is used to build a model that predicts a value or class for new cases. Common methods include regression, decision trees, random forests and neural networks. A model might, for instance, estimate a property price from floor area, location and year of construction. Before deployment it is tested on held-back data to detect overfitting.
Machine learning is a branch of artificial intelligence in which computers derive patterns from data instead of having every rule explicitly programmed. Depending on the task, models learn from labelled examples (supervised), from unlabelled data (unsupervised) or through rewards (reinforcement). Applications range from spam filters and demand forecasts to product recommendations. Quality depends heavily on the quantity and quality of the training data.
Data mining is the systematic search of large data sets for patterns, relationships and anomalies. It draws on statistical techniques and machine learning methods such as clustering, classification or association analysis. A classic example is market basket analysis in retail, which shows which products are often bought together – a basis for product range, placement and recommendations.
Exploratory data analysis (EDA) is the first, open-ended look at a dataset: you calculate summary figures, examine distributions and charts and look for patterns, outliers and missing values – without a fixed hypothesis yet. It helps assess data quality and find sensible questions for further analysis or modelling. Example: before a sales forecast, EDA reveals that one month was recorded incompletely because of a system migration.
