How to Generate a PDF Report

Options for different setups

Overview

There are several ways to turn your analysis into a polished PDF report. Which option works best depends on what software you already have installed. This guide covers six approaches, from the most feature-rich to the most minimal.

Option 1: Quarto (recommended if available)

Requires: Quarto (plus R or Python only if your .qmd executes those languages); pandoc is bundled with Quarto.

Quarto is the modern standard. It ships with pandoc, so no separate install is needed. You write code and narrative together in a .qmd file, and Quarto handles execution and typesetting.

Install:

# Windows (winget)
winget install --id Posit.Quarto --source winget --exact

# macOS (Homebrew)
brew install --cask quarto

# Or download from https://quarto.org/docs/get-started/

Verify setup:

quarto --version
quarto check

For PDF output, you also need LaTeX. The easiest route is TinyTeX:

quarto install tinytex

Render:

quarto render report.qmd              # uses format specified in YAML
quarto render report.qmd --to pdf     # force PDF output
quarto render report.qmd --to html    # HTML fallback (no LaTeX needed)

Tip: Quarto 1.4+ supports format: typst, which bundles its own typesetting engine and does not need LaTeX at all:

---
title: "My Report"
format: typst
---

Option 2: R Markdown

Requires: R + rmarkdown package + pandoc + LaTeX

The predecessor to Quarto. Uses .Rmd files with R code chunks. Very mature ecosystem with many templates.

If you do not use RStudio or Quarto, install pandoc separately (see Option 4).

Install:

# In R console
install.packages("rmarkdown")

# For LaTeX (minimal distribution, ~250 MB)
install.packages("tinytex")
tinytex::install_tinytex()

Verify setup:

rmarkdown::find_pandoc()
tinytex::is_tinytex()

Render:

rmarkdown::render("report.Rmd", output_format = "pdf_document")
rmarkdown::render("report.Rmd", output_format = "html_document")  # no LaTeX needed

Option 3: Python + LaTeX

Requires: Python (pandas, matplotlib) + xelatex or pdflatex

Python generates figures as PNGs and writes a .tex file. A LaTeX engine compiles it to PDF. This gives you full control over layout.

Install:

# Python packages
pip install pandas matplotlib numpy

# LaTeX -- pick one:
# Windows: MiKTeX (smaller/easier) or TeX Live
# https://miktex.org/download
# https://tug.org/texlive/
# macOS:
brew install --cask basictex     # minimal (~100 MB)
brew install --cask mactex       # full (~5 GB)
# Or use TinyTeX via R (see Option 2)

Verify setup:

xelatex --version   # or: pdflatex --version

Run:

python report.py
# Typical flow: your script writes report.tex, then you compile with xelatex

Manual compilation (if needed):

xelatex report.tex    # run twice for table of contents
xelatex report.tex

Option 4: Python + pandoc

Requires: Python (pandas, matplotlib) + pandoc + LaTeX

Python generates figures, you write the narrative in plain Markdown, and pandoc converts everything to PDF. Easier than writing raw LaTeX.

Install:

# Python packages
pip install pandas matplotlib numpy

# pandoc
# Windows:
winget install --source winget --exact --id JohnMacFarlane.Pandoc
# macOS:
brew install pandoc

# You still need LaTeX for PDF output (see Option 3)

Verify setup:

pandoc --version

Run:

# First generate your figures
python generate_figures.py

# Then convert Markdown to PDF
pandoc report.md -o report.pdf --pdf-engine=xelatex

# Or to HTML (no LaTeX needed)
pandoc report.md -o report.html --standalone

Option 5: Jupyter Notebook export

Requires: Python + Jupyter + nbconvert + pandoc + LaTeX (for PDF)

Write code and narrative together in .ipynb notebook cells, then export.

Install:

pip install jupyter nbconvert
# For PDF export, also install:
# - pandoc (see Option 4)
# - LaTeX (see Option 3)

Export:

jupyter nbconvert --to pdf notebook.ipynb     # needs pandoc + LaTeX
jupyter nbconvert --to html notebook.ipynb    # no LaTeX needed

Notebook UIs also offer Export/Download-as-PDF actions, but menu labels vary by app/version.

Option 6: Python + fpdf2 (no LaTeX needed)

Requires: Python 3.10+ (pandas, matplotlib, fpdf2)

A pure-Python option that produces text and figures in a PDF without any LaTeX or pandoc install. Limited typographic control but zero external dependencies.

Install:

pip install pandas matplotlib fpdf2

Verify setup:

python -c "from fpdf import FPDF; print('fpdf2 ok')"

Example:

from fpdf import FPDF

pdf = FPDF()
pdf.set_auto_page_break(auto=True, margin=15)

# Title page
pdf.add_page()
pdf.set_font("Helvetica", "B", 24)
pdf.cell(w=0, h=20, text="My Report Title", new_x="LMARGIN", new_y="NEXT", align="C")

# Section with text
pdf.set_font("Helvetica", "B", 14)
pdf.cell(w=0, h=12, text="1. Introduction", new_x="LMARGIN", new_y="NEXT")
pdf.set_font("Helvetica", size=11)
pdf.multi_cell(w=0, h=6, text="Your narrative text goes here...")

# Insert a figure
pdf.image("exhibits/fig1.png", w=170)

pdf.output("report.pdf")

Quick reference

Option	Tools needed	LaTeX required?	Notes
Quarto	Quarto (+ R/Python if executing code)	Yes (or use `typst`)	Best overall experience
R Markdown	R, rmarkdown, pandoc	Yes (use `tinytex`)	Mature, many templates
Python + LaTeX	Python, LaTeX engine	Yes	Full layout control
Python + pandoc	Python, pandoc	Yes (HTML fallback: no)	Write in Markdown
Jupyter export	Python, Jupyter, nbconvert, pandoc	Yes (HTML fallback: no)	Good for notebooks
Python + fpdf2	Python, fpdf2	No	Lowest barrier to PDF

Which should I pick?

Already have Quarto? Use Option 1. It’s the simplest end-to-end workflow.
Already have R? Use Option 2. Install tinytex for LaTeX and you’re set.
Only have Python + LaTeX? Use Option 3 (this is what we used for the earnings report).
Have Python but no LaTeX? Use Option 6 (fpdf2) for a quick PDF, or Option 4/5 with --to html as a fallback.
Want to avoid installing anything heavy? Use Quarto with format: typst (Option 1) or fpdf2 (Option 6) — neither needs LaTeX.

--- title: "How to Generate a PDF Report" subtitle: "Options for different setups" --- ## Overview There are several ways to turn your analysis into a polished PDF report. Which option works best depends on what software you already have installed. This guide covers six approaches, from the most feature-rich to the most minimal. ## Option 1: Quarto (recommended if available) **Requires:** Quarto (plus R or Python only if your `.qmd` executes those languages); pandoc is bundled with Quarto. Quarto is the modern standard. It ships with pandoc, so no separate install is needed. You write code and narrative together in a `.qmd` file, and Quarto handles execution and typesetting. **Install:** ```bash # Windows (winget) winget install --id Posit.Quarto --source winget --exact # macOS (Homebrew) brew install --cask quarto # Or download from https://quarto.org/docs/get-started/ ``` **Verify setup:** ```bash quarto --version quarto check ``` **For PDF output, you also need LaTeX.** The easiest route is TinyTeX: ```bash quarto install tinytex ``` **Render:** ```bash quarto render report.qmd # uses format specified in YAML quarto render report.qmd --to pdf # force PDF output quarto render report.qmd --to html # HTML fallback (no LaTeX needed) ``` **Tip:** Quarto 1.4+ supports `format: typst`, which bundles its own typesetting engine and does **not** need LaTeX at all: ```yaml --- title: "My Report" format: typst --- ``` ## Option 2: R Markdown **Requires:** R + `rmarkdown` package + pandoc + LaTeX The predecessor to Quarto. Uses `.Rmd` files with R code chunks. Very mature ecosystem with many templates. If you do not use RStudio or Quarto, install pandoc separately (see Option 4). **Install:** ```r # In R console install.packages("rmarkdown") # For LaTeX (minimal distribution, ~250 MB) install.packages("tinytex") tinytex::install_tinytex() ``` **Verify setup:** ```r rmarkdown::find_pandoc() tinytex::is_tinytex() ``` **Render:** ```r rmarkdown::render("report.Rmd", output_format = "pdf_document") rmarkdown::render("report.Rmd", output_format = "html_document") # no LaTeX needed ``` ## Option 3: Python + LaTeX **Requires:** Python (pandas, matplotlib) + xelatex or pdflatex Python generates figures as PNGs and writes a `.tex` file. A LaTeX engine compiles it to PDF. This gives you full control over layout. **Install:** ```bash # Python packages pip install pandas matplotlib numpy # LaTeX -- pick one: # Windows: MiKTeX (smaller/easier) or TeX Live # https://miktex.org/download # https://tug.org/texlive/ # macOS: brew install --cask basictex # minimal (~100 MB) brew install --cask mactex # full (~5 GB) # Or use TinyTeX via R (see Option 2) ``` **Verify setup:** ```bash xelatex --version # or: pdflatex --version ``` **Run:** ```bash python report.py # Typical flow: your script writes report.tex, then you compile with xelatex ``` **Manual compilation (if needed):** ```bash xelatex report.tex # run twice for table of contents xelatex report.tex ``` ## Option 4: Python + pandoc **Requires:** Python (pandas, matplotlib) + pandoc + LaTeX Python generates figures, you write the narrative in plain Markdown, and pandoc converts everything to PDF. Easier than writing raw LaTeX. **Install:** ```bash # Python packages pip install pandas matplotlib numpy # pandoc # Windows: winget install --source winget --exact --id JohnMacFarlane.Pandoc # macOS: brew install pandoc # You still need LaTeX for PDF output (see Option 3) ``` **Verify setup:** ```bash pandoc --version ``` **Run:** ```bash # First generate your figures python generate_figures.py # Then convert Markdown to PDF pandoc report.md -o report.pdf --pdf-engine=xelatex # Or to HTML (no LaTeX needed) pandoc report.md -o report.html --standalone ``` ## Option 5: Jupyter Notebook export **Requires:** Python + Jupyter + nbconvert + pandoc + LaTeX (for PDF) Write code and narrative together in `.ipynb` notebook cells, then export. **Install:** ```bash pip install jupyter nbconvert # For PDF export, also install: # - pandoc (see Option 4) # - LaTeX (see Option 3) ``` **Export:** ```bash jupyter nbconvert --to pdf notebook.ipynb # needs pandoc + LaTeX jupyter nbconvert --to html notebook.ipynb # no LaTeX needed ``` Notebook UIs also offer Export/Download-as-PDF actions, but menu labels vary by app/version. ## Option 6: Python + fpdf2 (no LaTeX needed) **Requires:** Python 3.10+ (pandas, matplotlib, fpdf2) A pure-Python option that produces text and figures in a PDF without any LaTeX or pandoc install. Limited typographic control but zero external dependencies. **Install:** ```bash pip install pandas matplotlib fpdf2 ``` **Verify setup:** ```bash python -c "from fpdf import FPDF; print('fpdf2 ok')" ``` **Example:** ```python from fpdf import FPDF pdf = FPDF() pdf.set_auto_page_break(auto=True, margin=15) # Title page pdf.add_page() pdf.set_font("Helvetica", "B", 24) pdf.cell(w=0, h=20, text="My Report Title", new_x="LMARGIN", new_y="NEXT", align="C") # Section with text pdf.set_font("Helvetica", "B", 14) pdf.cell(w=0, h=12, text="1. Introduction", new_x="LMARGIN", new_y="NEXT") pdf.set_font("Helvetica", size=11) pdf.multi_cell(w=0, h=6, text="Your narrative text goes here...") # Insert a figure pdf.image("exhibits/fig1.png", w=170) pdf.output("report.pdf") ``` ## Quick reference | Option | Tools needed | LaTeX required? | Notes | |--------|-------------|-----------------|-------| | **Quarto** | Quarto (+ R/Python if executing code) | Yes (or use `typst`) | Best overall experience | | **R Markdown** | R, rmarkdown, pandoc | Yes (use `tinytex`) | Mature, many templates | | **Python + LaTeX** | Python, LaTeX engine | Yes | Full layout control | | **Python + pandoc** | Python, pandoc | Yes (HTML fallback: no) | Write in Markdown | | **Jupyter export** | Python, Jupyter, nbconvert, pandoc | Yes (HTML fallback: no) | Good for notebooks | | **Python + fpdf2** | Python, fpdf2 | **No** | Lowest barrier to PDF | ## Which should I pick? - **Already have Quarto?** Use Option 1. It's the simplest end-to-end workflow. - **Already have R?** Use Option 2. Install `tinytex` for LaTeX and you're set. - **Only have Python + LaTeX?** Use Option 3 (this is what we used for the earnings report). - **Have Python but no LaTeX?** Use Option 6 (`fpdf2`) for a quick PDF, or Option 4/5 with `--to html` as a fallback. - **Want to avoid installing anything heavy?** Use Quarto with `format: typst` (Option 1) or `fpdf2` (Option 6) — neither needs LaTeX.