Statistics is the discipline that transforms raw observations into structured knowledge through systematic collection, cleaning, analysis, and communication of data. It formalises uncertainty with probability, provides tools to separate signal from noise, and links measured evidence to actionable conclusions. Statistical thinking emphasises transparent assumptions, reproducible procedures, and careful interpretation so that results carry appropriate confidence when applied beyond the observed sample.
Descriptive statistics summarise data with measures of centre such as the mean, median, and mode, measures of spread including variance, standard deviation, and interquartile range, and measures of shape like skewness and kurtosis. Visual summaries such as histograms, box plots, density plots, and scatterplots reveal distributional features and relationships at a glance. Descriptive methods also include tabulation, aggregation, and smoothing techniques that prepare data for further modeling while highlighting outliers and data quality issues.
Inferential statistics use sample information to make probabilistic statements about wider populations with tools for estimation and hypothesis testing. Confidence intervals quantify plausible ranges for parameters and p values measure evidence against null hypotheses. Sampling design, bias assessment, and power calculations are central to valid inference. Modern inferential practice integrates model checking, sensitivity analysis, and resampling methods that improve robustness when classical assumptions are strained.
Statistical modeling builds simplified representations of relationships between variables using regression, generalized linear models, time series methods, and nonparametric approaches. Model selection, cross validation, regularisation, and assessment of predictive performance guard against overfitting. Probabilistic models deliver calibrated uncertainty for forecasts and decision rules that trade off risks, costs, and benefits.
Statistics underpins decision making across science, public policy, business, and everyday life by converting data into evidence that acknowledges uncertainty. Responsible practice demands transparent reporting, attention to sampling and measurement error, avoidance of data dredging, and explicit discussion of limitations. Ethical stewardship of data includes protecting privacy, disclosing conflicts, and ensuring that statistical conclusions are not misrepresented.