Key Terms and Coding Vocabulary

This is a learning-support page, not just a dictionary. Terms are grouped by topic. Each entry gives a plain-language definition, a small forestry example, a short code example where it helps, one common mistake, and often a check your understanding question. Throughout the book, the first time a hard term appears in a chapter it links here.

Tip

Use the table of contents (or your browser’s find, Ctrl/Cmd + F) to jump to a term. Every term heading is a link target.

Excel and data basics

Data dictionary

A short table — one row per column — recording each variable’s name, meaning, units, and type. Forestry example: for the BC silviculture file, one row says Reforestation_ha = “area reforested that fiscal year, in hectares, numeric.” Common mistake: writing the dictionary from memory instead of checking the file’s About/ReadMe sheets. Check: what three things should every dictionary row record?

Cleaning log

A running record of every change you make to raw data and why, so another analyst can reproduce your cleaned file. Forestry example: “Dropped 12 all-missing pollutant columns because no station measured them this month.” Common mistake: cleaning data with clicks and never writing down what you did.

Cell reference (relative vs absolute)

How an Excel formula points at other cells. A relative reference (H2) shifts when copied down; an absolute reference ($H$2) stays fixed. Common mistake: forgetting the $ in a lookup range, so the range “slides” as you copy the formula and returns wrong answers. Check: which reference stays put when you copy a formula, G2 or $G$2?

Structured table

An Excel range converted to a named table (Ctrl+T) so sorting, filtering, and formulas stay safe as data grows. Common mistake: sorting one column by itself and scrambling the rows — a structured table sorts all columns together.

PivotTable

An Excel tool that summarises a table by dragging fields into Rows, Columns, Values, and Filters — the spreadsheet equivalent of group_by() + summarise(). Forestry example: mean harvested area by decade; total area by forest type × seral stage.

Missing value

A value that is not recorded. In Excel it is a blank cell; in R it is NA. It is not zero and not an empty string — it means “we do not know.” Forestry example: a station with no ozone sensor has NA for O₃, which is different from a real reading of 0 ppb. Common mistake: letting NA be treated as 0, which silently corrupts every sum and mean. Use mean(x, na.rm = TRUE). Check: is a missing DBH the same as a DBH of 0?

Open Government Licence (OGL)

The licence under which BC, Ontario, and federal open data are released, permitting reuse with attribution. Every dataset in this book is open data.

R basics

Object

Anything you store in R with a name using <- (data, a number, a plot).

bc <- read_excel(here("data", "bc_disturbance_reforestation.xlsx"))

Here bc is an object. Common mistake: using an object before you have created it → object 'bc' not found.

Vector

An ordered set of values of the same type — R’s most basic data container. A column of a table is a vector.

ha <- c(220000, 180000, 250000)   # a numeric vector

Check: can one vector hold both numbers and text at the same time?

Function

A named command that does something to its inputs and returns a result. You call it with parentheses: mean(x). Forestry example: sum(bc$Reforestation_ha) adds up a column. Common mistake: forgetting the parentheses (mean vs mean(x)).

Argument

An input you give a function, inside its parentheses. In read_excel(path, sheet = "Data"), path and sheet are arguments. Common mistake: misspelling an argument name (sheets = instead of sheet =) or passing arguments in the wrong order.

Package

A bundle of extra functions you install once and load each session with library().

install.packages("readxl")   # once
library(readxl)              # every session

Common mistake: could not find function "read_excel" — the package is installed but you forgot library(readxl).

Working directory

The folder R treats as “here” when it looks for files. In this course you never set it by hand — the here package finds the project root for you.

here("data", "airdata3.csv")   # a portable path

Common mistake: hard-coding C:/Users/you/Desktop/..., which breaks on every other computer.

Data frame

R’s classic rectangular table: rows are observations, columns are variables. See also: Tibble, the tidyverse’s tidier data frame.

Tibble

A modern data frame used by the tidyverse. It prints a compact preview, shows column types, and never silently changes your data. Forestry example: read_csv("airdata3.csv") returns a tibble of hourly air-quality readings. Common mistake: expecting a tibble to behave exactly like a spreadsheet — you change it with code (mutate, filter), not by clicking cells. Check: how does a tibble differ from an Excel sheet?

Pipe

The native operator |> that passes the result on its left into the function on its right, so a workflow reads left-to-right.

bc |> filter(Fiscal_Year >= 2000) |> summarise(mean(Reforestation_ha))

Read |> as “and then.” Common mistake: starting a new line with |> at the front instead of ending the previous line with it.

Tidyverse vocabulary

Filter

Keep or drop rows by a condition. filter() ends in “R” — think Rows.

airdata |> filter(!is.na(O3))     # keep rows where O3 was measured

Common mistake: using filter() when you meant select() (columns).

Select

Keep or drop columns by name. select() ends in “C” — think Columns.

bc |> select(Fiscal_Year, Reforestation_ha)

Mutate

Add or change a column. The number of rows stays the same.

bc |> mutate(decade = floor(Fiscal_Year / 10) * 10)

Common mistake: confusing mutate() (adds columns, same rows) with summarise() (collapses to fewer rows).

Summarize

Collapse many rows into summary values (a mean, a count). Usually paired with group_by().

airdata |> group_by(location) |> summarise(mean_o3 = mean(O3, na.rm = TRUE))

Common mistake: forgetting na.rm = TRUE, so one NA makes the whole summary NA. (Spelled summarise or summarize — both work.)

group_by

Split a table into groups so the next summarise() (or mutate()) runs per group. Forestry example: group_by(region) then summarise() gives one row per region. Common mistake: leaving a table grouped after you are done — add .groups = "drop" or ungroup().

Join

Combine two tables by a matching column. left_join() keeps every row of the left table and attaches matching columns from the right.

airdata |> left_join(station_regions, by = "location")

Common mistake: not checking row counts before and after — a duplicated key can multiply your rows. Check: after a left_join, should the left table have more rows than before?

Key

The column(s) two tables are joined on (e.g., location). A good key is unique in at least one table. Common mistake: joining on a key that repeats in both tables, creating an unexpected many-to-many explosion.

Pivot

Change a table’s shape. pivot_longer() folds many columns into key/value rows; pivot_wider() spreads a key column across new columns. Forestry example: turn a wide table (one column per year) into a long table (one row per year) so it is easy to plot. Common mistake: reshaping when you did not need to — pivot only when the current shape can’t answer the question.

Quarto and reproducible reporting

Quarto

The open-source publishing system that turns a .qmd file (prose + code + output) into an HTML report, PDF, Word document, or slides. Forestry example: your lab report, its code, and its figures all live in one .qmd.

Chunk

A block of code inside a .qmd, fenced by ```{r} … ```. When you render, R runs the chunk and inserts its output. Common mistake: code that works in the console but errors on render because an earlier chunk (that created an object) was skipped.

Render

The act of running a .qmd from top to bottom to produce the finished document. In RStudio, the Render button. Common mistake: assuming a report is reproducible without rendering it from a clean session — always render before submitting.

Reproducible workflow

An analysis another person (or future you) can re-run from the raw data to the same result, because every step is recorded in code. Check: if you deleted all your outputs, could you regenerate them by pressing Render?

Visualization

Visualization

A chart or map that turns numbers into a picture so patterns are easy to see and communicate. Forestry example: a line chart of reforested area by fiscal year.

Geom

A ggplot2 geometry — the kind of mark a layer draws: geom_point(), geom_line(), geom_col(), geom_boxplot(). Common mistake: using geom_bar() (which counts rows) when you have pre-computed values and need geom_col().

Aesthetic mapping

The aes() part of a plot: which columns are shown by which visual properties (x-axis, y-axis, colour, fill, size, shape).

ggplot(bc, aes(x = Fiscal_Year, y = Reforestation_ha))

Common mistake: putting a constant colour inside aes() (which makes a misleading legend) instead of outside it.

Facet

Splitting one plot into small multiples — one panel per group — with facet_wrap(). Forestry example: one ozone time-series panel per monitoring region.

AI-assisted debugging

Prompt

The message you give an AI assistant. A good prompt gives context + dataset + task + what you tried + the error/output + what you need (see How to use this book → How to write a useful AI prompt).

Hallucination

When an AI invents something that looks right but is false — a function that does not exist, a column you never had, or a made-up number. Common mistake: pasting AI code that calls an invented function and trusting it. Check functions against the package documentation. Check: how would you catch an AI that invented a column name?

Verification

Confirming a result is correct yourself — with row counts, missing-value checks, summary statistics, and visual inspection — rather than trusting code (yours or an AI’s) on faith. Verification is the habit this whole book is built around.

Spatial terms (optional extension)

CRS (coordinate reference system)

The system that ties spatial coordinates to locations on Earth (e.g., WGS 84, BC Albers). Reprojecting converts data from one CRS to another. Covered in the optional Spatial Data Analysis in R resource.