CODING
Instructions and environments.
- Object
- A named value stored in an R session, such as a number, table, or model.
- Function
- A reusable operation that takes inputs, performs work, and returns an output.
- Argument
- A value supplied to a function to control what it does.
- Pipe
- The
|>operator that passes the result of one step into the next step. - Script
- A plain-text file of code that can be saved, rerun, reviewed, and shared.
- Console
- The place where R executes commands immediately. Console history is not a reproducible script.
- Working directory
- The folder R treats as its current location when resolving relative paths.
- Clean session
- A newly restarted R process with no objects left from earlier work.
- Render
- Build a source document, such as Quarto, into HTML, PDF, or slides.
DATA
Rows, variables, and origins.
- Unit of observation
- The entity represented by one row. Here, the default unit is usually a country-year.
- Variable
- A measured or constructed attribute stored in a column.
- Key
- One or more variables whose values uniquely identify each row.
- Missing value
- An observation that is unavailable or not recorded. In R it is commonly represented as
NA. - Tidy data
- A structure where each variable is a column, each observation is a row, and each type of unit is a table.
- Join
- An operation that combines tables by matching keys.
- Provenance
- The documented origin and processing history of data.
- Indicator
- A defined measure used to represent a condition or change, often published with a code and unit.
STATISTICS
Patterns and their limits.
- Distribution
- The range, frequency, and shape of observed values.
- Transformation
- A systematic change to a variable, such as taking a logarithm, to change scale or interpretation.
- Regression
- A model that summarizes how an outcome varies with one or more predictors.
- Coefficient
- A model quantity that describes an estimated change in the outcome associated with a change in a predictor, holding included predictors constant.
- Residual
- The difference between an observed outcome and the model’s fitted value.
- Association
- A pattern in which variables vary together. Association alone does not establish causation.
- Causal effect
- The change in an outcome produced by an intervention or exposure. Identifying it requires a research design beyond a descriptive regression.
- Analysis sample
- The observations actually used in a plot, summary, or model after filters and missingness.
AI AND AGENTS
Generated help and human checks.
- Generative AI
- A model that produces text, code, images, or other content from instructions and context.
- Coding agent
- An AI tool that can inspect a project, propose file changes, run commands, and report results.
- Prompt
- The instruction and context given to a model or agent.
- Context
- The files, conversation, constraints, and other information available to the agent for a task.
- Hallucination
- Generated content that sounds plausible but is unsupported, false, or invented.
- Diff
- A line-by-line representation of what changed in a file.
- Verification
- An independent check of a claim using code behavior, data, output, documentation, or a primary source.
- AI-use note
- A brief record of what an AI tool contributed and how the human checked it.