Reference

Glossary

---

Automation

Replacing manual, interactive steps with repeatable scripted processes to enable consistent and scalable analysis.

Big Data

Data becomes “big” when your normal way of working breaks, not simply when it reaches large size.

Big Data Thinking

Adapting workflows, tools, and research approaches when traditional methods no longer scale.

Caching

Saving intermediate results so that only steps with changed inputs need to be re-run.

Data Provenance

A record of where data comes from, including its source, version, selection criteria, and any processing or transformations applied.

Data Reduction

The process of cleaning, filtering, calibrating, or transforming raw data into a form suitable for analysis.

Dependency (Workflow)

A relationship where one step in a workflow requires the output of another step before it can run.

Dependency (Software)

The set of software tools or libraries that are required to perform a task, or in order to use another set of tools or libraries.

Distributed Computing

Computation and data storage spread across multiple machines or locations.

Documentation

Written descriptions of data, methods, assumptions, and limitations that enable understanding and reproducibility.

FAIR Principles

Guidelines for making data and software Findable, Accessible, Interoperable, and Reusable through clear documentation and sharing practices.

High-Performance Computing (HPC)

Specialised computing systems designed to process large datasets or complex workflows using parallel resources.

Modularity

Designing workflows as independent components that can be reused, modified, or replaced without affecting the entire system.

Parallelism

Running independent tasks in a workflow simultaneously to reduce total execution time.

Pipeline (or Workflow)

A sequence of automated processing steps that transforms raw data into final results or science-ready outputs.

README

A short document describing a project, including its purpose, data sources, and how to reproduce key results.

Reproducible Research

Research where the main results can be rerun and others can reproduce the analysis without guessing missing steps.

Reuse

Using existing data, code, or workflows for new analyses.

Streaming Data

Data that arrives continuously over time, requiring real-time or near-real-time processing.

Versioning

Tracking changes to data, code, or workflows over time so results can be reproduced and compared.

Workflow

A structured description of the steps in a research process, including inputs, methods, tools, and outputs.

Workflow (Computational)

An automated sequence of data processing steps implemented as scripts or programs that reproduce a research process.

Workflow (Human)

A conceptual description of a process (like a recipe) that can be followed step-by-step by a person to achieve a result.

Workflow Manager

A tool that organises, executes, and tracks workflows, often handling dependencies, parallel execution, and reproducibility.