Activity 07 — Your First Quarto Report

Read the data, plot it, describe it, test it — then render to Word, PDF, and slides

quarto

reproducibility

markdown

Follow-along activity paralleling the Quarto lecture. Students fill in a starter .qmd that takes the leaf data from start to finish — read it in, box plot, mean ± SE plot, histogram, summary statistics table, and a Welch’s t-test — and render it to Word, PDF, and RevealJS slides.

Author

Bill Perry

Published

October 1, 2026

How this activity works

You are building one Quarto file this whole class, and you turn it in: leaf_report.qmd.

Everything you have done so far — reading data, plotting, summary statistics — goes into one document that also holds your writing.

The analysis code is already in the starter file, and it runs. You add the finishing touches (colors, labels, theme, table formatting), write the text, and render. Whatever time is left is for Part 9 — play.

Start every code chunk with a short # comment that says what it does. Same rule as your scripts; it is part of the grade.

Render often — after every chunk. A break is easy to find when you only changed one thing.

🔮 Predict before you Render

Before each Render, say what the document will show — a heading, a plot, a table, a number. Then check.

📥 Get the files

Starter file → leaf_report_skeleton.qmd — save it at the top of your project folder as leaf_report.qmd. It already has the YAML, every section heading, and working code for every step — plots, summary table, and t-test — ready for you to finish.
Leaf data → 2026_09_03_data_sci_leaf_area.xlsx — should already be in your data/ folder.
r_themes_for_3_sizes.R and summary_stats_function.R — should already be in your themes/ folder (Activities 04 and 06).

What you turn in (Part 8): leaf_report.qmd, leaf_report.docx, and leaf_report.pdf.

Part 1 · Put the file in the right place, then Render it

Save the starter file at the top level of your project — the same folder that holds data/ and themes/:

leaf_project/
├── data/
│   └── 2026_09_03_data_sci_leaf_area.xlsx
├── figures/
├── scripts/
├── themes/
│   ├── r_themes_for_3_sizes.R
│   └── summary_stats_function.R
└── leaf_report.qmd          <- HERE, at the top

Why the top of the project?

Quarto runs your code from the folder the .qmd is saved in. Saved at the top of the project, "data/2026_09_03_data_sci_leaf_area.xlsx" and "themes/..." work exactly as they do in your scripts — no here package, no ../.

If you saved it inside documents/ instead, every path would need ../data/... to climb back up a folder. Keep it simple: top of the project.

Open leaf_report.qmd in Positron and press Render (Ctrl/Cmd + Shift + K) before you change anything. You should get a document with a title, headings, three plain plots, two plain tables, and a t-test printout. It all works — it just isn’t finished.

Did it render?  Y / N
If not, what did the error message name?

Part 2 · The YAML — one switch for all the code

Look at the block between the two --- lines at the top. Change author: to your name.

title: "Leaf Mass on the Sunny and Shady Sides of a Tree"
author: "Your Name"
date: today

format:
  docx:
    toc: true
    number-sections: true
  pdf:
    toc: true
    number-sections: true
    fig-pos: "H"     # keep each figure right under its heading
  revealjs:
    scrollable: true
    smaller: true

execute:
  echo: true       # false = hide ALL the code
  warning: false
  message: false

format: lists three outputs — Word, PDF, and RevealJS slides — from the same file.
execute: sets the default for every chunk. echo: true shows the code; echo: false hides all of it at once.

🔮 Predict: If you change echo: true to echo: false and Render, what disappears from the document, and what stays?

✏️ Your turn: Make the change, Render, look — then set it back to true for now. You will turn it off for good in Part 8.

What disappeared:
What stayed:

Warning

YAML is space-sensitive. Indent with two spaces, never tabs. If Render fails right away, the YAML is the first place to look.

Part 3 · Markdown — write the Introduction and Methods

Replace the  under Introduction with two or three sentences, and fill in the hypotheses:

# Introduction

Leaves on the **shady** side of a tree may grow larger and
heavier to catch more of the limited light. We collected leaves
from the sunny and shady sides of a tree and weighed each one.

- **Null hypothesis:** mean leaf mass is the same on both sides
- **Alternate hypothesis:** mean leaf mass differs between sides

✏️ Your turn: Write two or three sentences under Methods in your own words: how the leaves were weighed, which plots and summary statistics you made, and that you compared the sides with a Welch’s two-sample t-test. Render.

Note

 is a Markdown comment — it never shows up in the rendered document. The skeleton uses them as notes to you. Delete them as you fill each section in.

Part 4 · Chunk options — your reference card

Every chunk in the skeleton starts with lines that begin #|. Those are chunk options. Keep this table handy — these are the ones you will use all term:

Common code-chunk options. The `execute:` block in the YAML sets the default; a chunk’s own `#|` line overrides it for that chunk only.
Option	What it does	Example
`label`	Names the chunk. `fig-` or `tbl-` at the front numbers it as a Figure or Table	`#\| label: fig-boxplot`
`echo`	Show (`true`) or hide (`false`) the code	`#\| echo: false`
`eval`	Run the code (`true`) or only show it (`false`)	`#\| eval: false`
`include`	`false` runs the code but hides both code and output	`#\| include: false`
`output`	`false` hides the results but keeps the code	`#\| output: false`
`warning`	Show or hide warnings	`#\| warning: false`
`message`	Show or hide package start-up messages	`#\| message: false`
`fig-width`	Figure width in inches	`#\| fig-width: 5`
`fig-height`	Figure height in inches	`#\| fig-height: 4`
`fig-cap`	Figure caption	`#\| fig-cap: "Leaf mass by side"`
`tbl-cap`	Table caption	`#\| tbl-cap: "Summary statistics"`
`tbl-colwidths`	Column widths (percent) so Word tables line up	`#\| tbl-colwidths: [30, 70]`

Look at the first two chunks in the skeleton — they are already filled in:

```{r}
#| label: setup
# load packages ----
library(readxl)      # read Excel files
library(tidyverse)   # dplyr + ggplot2
library(janitor)     # clean_names()
library(broom)       # tidy() turns test output into a table

# load our theme file and summary_stats() helper ----
source("themes/r_themes_for_3_sizes.R")
source("themes/summary_stats_function.R")
```

```{r}
#| label: load-data
# read the leaf data from the data/ folder ----
leaf_df <-
  read_excel("data/2026_09_03_data_sci_leaf_area.xlsx") %>%
  clean_names()
```

🔮 Predict: The setup chunk prints nothing useful to a reader. Which one option from the table would make it run but leave no trace in the document?

✏️ Your turn: Add that option to the setup chunk and Render. Did the library code disappear?

Option I added:
Did the packages still load (do later chunks still work)?  Y / N

Tip

Labels must be unique. Two chunks named plot stop the Render with “Duplicate chunk label”. The error names the label — go straight to it.

Part 5 · Finish the plots

The three plots are already in your skeleton, and they run. Render and look at them. They are correct, but they are not ready for a report: default colors, axis labels that read mass_g and shade, and a legend that repeats what the x-axis already says.

Your job is the finishing touches — the same layers you learned in Activities 03–04. Each chunk ends with a # TODO: comment showing where they go.

🔮 Predict: Before you change anything, list three things about the plain box plot a reader would find confusing.

1.
2.
3.

Box plot

Here is the finished chunk. The last four lines are yours to add in place of the # TODO: comment:

```{r}
#| label: fig-boxplot
#| fig-cap: "Leaf mass by side; each point is one leaf."
#| fig-width: 5
#| fig-height: 4
# box plot with raw points on top ----
leaf_df %>%
  ggplot(aes(x = shade, y = mass_g, fill = shade)) +
  geom_boxplot(alpha = 0.6, outlier.shape = NA) +
  geom_jitter(width = 0.15, alpha = 0.6) +
  scale_fill_manual(values = c("sunny" = "goldenrod",
                               "shady" = "forestgreen")) +
  labs(x = "Side of tree", y = "Leaf mass (g)") +
  theme_regular() +
  theme(legend.position = "none")
```

Warning

Adding a layer means adding a + to the end of the line above it. Forget it and R runs the plot without your new lines — or stops with an error.

Render. Notice the caption starts “Figure 1:” — that number came from the fig- at the front of the label.

Mean ± SE — your turn

✏️ Add the same four finishing lines to the fig-mean-se chunk. Careful: this plot maps color = shade, not fill = shade. Which scale_..._manual() do you need?

Check your answer

```{r}
#| label: fig-mean-se
#| fig-cap: "Mean leaf mass (± 1 SE) by side."
#| fig-width: 5
#| fig-height: 4
# mean +/- standard error plot ----
leaf_df %>%
  ggplot(aes(x = shade, y = mass_g, color = shade)) +
  geom_jitter(width = 0.15, alpha = 0.35, size = 2) +
  stat_summary(fun = mean, geom = "point", size = 4) +
  stat_summary(fun.data = mean_se, geom = "errorbar",
               width = 0.15, linewidth = 0.9) +
  scale_color_manual(values = c("sunny" = "goldenrod",
                                "shady" = "forestgreen")) +
  labs(x = "Side of tree", y = "Leaf mass (g)") +
  theme_regular() +
  theme(legend.position = "none")
```

Histogram — your turn

✏️ Finish the fig-histogram chunk on your own. This time the x-axis is leaf mass and the y-axis is a count.

Check your answer

```{r}
#| label: fig-histogram
#| fig-cap: "Distribution of leaf mass on each side."
#| fig-width: 6
#| fig-height: 3.5
# one histogram per side ----
leaf_df %>%
  ggplot(aes(x = mass_g, fill = shade)) +
  geom_histogram(binwidth = 0.05, color = "white") +
  facet_wrap(~shade) +
  scale_fill_manual(values = c("sunny" = "goldenrod",
                               "shady" = "forestgreen")) +
  labs(x = "Leaf mass (g)", y = "Count") +
  theme_regular() +
  theme(legend.position = "none")
```

✏️ Your turn: Change fig-width on the histogram from 6 to 3 and Render. Then put it back.

What happened to the histogram at fig-width: 3?
Looking at all three plots: which side looks heavier, if either?

Part 6 · Finish the summary statistics table

The table chunk is already in your skeleton — the same summary_stats() helper from Activity 06, printed with knitr::kable(), which turns a data frame into a real table in Word, PDF, and slides. Render and look: seven decimal places and headers like ci_lower.

✏️ Finish it by filling in kable():

```{r}
#| label: tbl-summary
#| tbl-cap: "Summary statistics for leaf mass (g) by side."
#| tbl-colwidths: [12, 7, 12, 15, 12, 12, 15, 15]
# summary statistics by side, printed as a table ----
mass_stats_df <- leaf_df %>%
  group_by(shade) %>%
  summary_stats(mass_g)

mass_stats_df %>%
  knitr::kable(
    digits = 3,
    col.names = c("Side", "n", "Mean", "Variance",
                  "SD", "SE", "CI low", "CI high"))
```

digits = 3 rounds every number in the table — no round() needed
col.names = gives the columns readable names (one per column, in order)
tbl-colwidths: (already in the skeleton) sets each column’s width in percent — eight columns, eight numbers. Without it, Word guesses and often squeezes the last column.

Under the table, point the reader to it — let Quarto fill in the number:

@tbl-summary gives the sample size, mean, variance, standard
deviation, standard error, and 95% confidence interval for leaf
mass on each side.

Render. @tbl-summary turns into “Table 1” — and stays correct if you ever add a table above it.

Mean mass, shady:          sunny:
Which side has the larger SD?

Part 7 · Test it — Welch’s two-sample t-test

You met the t-test in introductory biology: does the difference between two means look bigger than chance? We use Welch’s version, which does not assume the two sides have equal variances — that is what var.equal = FALSE says. (Next lecture takes the t-test apart properly.)

The test is already in your skeleton:

```{r}
#| label: ttest
# Welch's t-test (does NOT assume equal variances) ----
leaf_ttest <- t.test(mass_g ~ shade, data = leaf_df,
                     var.equal = FALSE)
leaf_ttest
```

mass_g ~ shade reads “mass by shade” — the number on the left, the two groups on the right
The printout says Welch Two Sample t-test at the top — check that it does

The printout is fine for you, but ugly in a report. The next chunk uses tidy() from broom to turn it into a data frame, so kable() can make it a table.

✏️ Your turn: Finish the tbl-ttest chunk the same way you finished Part 6 — digits = 3 and six readable col.names.

Check your answer

```{r}
#| label: tbl-ttest
#| tbl-cap: "Welch's t-test of leaf mass by side of the tree."
#| tbl-colwidths: [20, 20, 15, 15, 15, 15]
# the same result as a clean table ----
leaf_ttest %>%
  tidy() %>%
  select(estimate1, estimate2, statistic, parameter,
         p.value, conf.low) %>%
  knitr::kable(
    digits = 3,
    col.names = c("Mean shady", "Mean sunny", "t",
                  "df", "p-value", "CI low"))
```

🔮 Predict before you Render: Look back at your mean ± SE plot. Do the error bars overlap a lot or a little? Will the p-value be below 0.05?

My prediction:
t =        df =        p-value =
Is p below 0.05?  Y / N

✏️ Your turn — Discussion: Under Discussion, write two or three sentences. Point to @fig-mean-se and @tbl-ttest, say which side had heavier leaves and by how much, and whether the p-value is below 0.05. Do the data support the alternate hypothesis? Refer to the tables — don’t type the numbers in by hand.

Note

Our leaves give a p-value well above 0.05, so we fail to reject the null hypothesis. That is a real result, not a failed analysis — report it honestly.

Part 8 · Hide the code and render all three formats

A report is for a reader who wants results, not R. Go back to the YAML and make the one-line change:

execute:
  echo: false      # hide ALL the code

One more tidy-up: the raw t.test() printout was for you. Now that @tbl-ttest holds the same numbers, hide the printout but keep the chunk running (it still has to make leaf_ttest). From your reference card:

```{r}
#| label: ttest
#| output: false
# Welch's t-test (does NOT assume equal variances) ----
leaf_ttest <- t.test(mass_g ~ shade, data = leaf_df,
                     var.equal = FALSE)
leaf_ttest
```

The Render button in Positron previews one format. To build all three outputs listed under format:, open the Terminal tab (not the Console) and run:

quarto render leaf_report.qmd

You should now have three new files next to your .qmd:

The three outputs from one `.qmd`.
File	What it is
`leaf_report.docx`	Word report — hand this in
`leaf_report.pdf`	PDF report — hand this in
`leaf_report.html`	RevealJS slide show — open it in a browser, press F for full screen and the arrow keys to move

Note

When Word opens leaf_report.docx it may ask to update fields — say Yes. That fills in the table of contents.

PDF says it needs LaTeX?

PDF output needs a small LaTeX install, once per computer. In the Terminal:

quarto install tinytex

Then run quarto render leaf_report.qmd again.

✏️ Your turn: Open the slides. Each # heading became a section slide and each ## heading became its own slide — the same file is now a presentation.

How many slides did you get?
Which slide would you show first in lab meeting?

Part 9 · Play — if you finish early

Render and hand in first. Then File → Save As leaf_report_play.qmd so your hand-in stays safe, and try any of these in the copy:

A different variable. Change mass_g to petiole_mm in every chunk (Ctrl/Cmd + F to find them all), fix the axis labels and captions, and Render. The whole report — plots, tables, test — rebuilds for the new variable.
A new figure. Add a chunk with #| label: fig-scatter and a fig-cap, plot mass_g against petiole_mm with geom_point(), and refer to it in the text with @fig-scatter.
Theme sizes. Swap theme_regular() for theme_small() or theme_large(). Which one fits a 5-inch-wide figure best?
Slide style. Under revealjs: in the YAML add theme: dark (or sky, serif, solarized) and render the slides again.
Chunk options. Put #| echo: true on just one chunk while the YAML says echo: false. What shows up?

🧰 Keep this file

leaf_report.qmd is now your template for every analysis this term — the t-test activities next week, and your final project. Copy it, change the data file in load-data and the column names in the chunks, rewrite the text, and Render.

Check yourself

You are done when your report has:

Your name in the YAML, and echo: false so no code shows
An Introduction with both hypotheses, and a Methods paragraph
A box plot, a mean ± SE plot, and a histogram — each with your colors, real axis labels, theme_regular(), no legend, and a numbered caption
A summary statistics table with readable column names and even column widths
The Welch’s t-test as a table (raw printout hidden with output: false)
A Discussion that points to the figures and tables with @fig-... / @tbl-...
Every # TODO: comment replaced by your code

Turn it in

Submit all three:

leaf_report.qmd — your source file
leaf_report.docx — the Word report
leaf_report.pdf — the PDF report

References: 📖 R4DS Ch 28 — Quarto · 📖 R4DS Ch 29 — Formats · 🌐 Quarto code-chunk options · 🌐 Quarto RevealJS