Week 3: Publishing — GitHub Pages, LaTeX, and Pull Requests

Week 3: Publishing — GitHub Pages, LaTeX, and Pull Requests#

A pipeline’s output has to land somewhere people can see it. Last week the pipeline learned to build its own notebooks and its own ChartBook site. This week is the publication end: GitHub Pages as the host for that site, and LaTeX for the PDF. The habit to build is treating a report not as a document you write once but as a build target the pipeline regenerates whenever the data changes.

Announcements#

  • HW 1 is due next Tuesday, October 20, at the start of class. This is the week to finish it: ask your questions tonight, because from here on the material is HW 2.

  • HW 2 launches today, due Tuesday, October 27, two weeks from tonight: the Treasury yield curve and the policy path, published as a ChartBook site on GitHub Pages and summarized in a PDF market brief.

  • Project assignments. Each group is emailed its assigned paper from the Potential Final Projects list. If you have not submitted your preferences, do it today.

  • Book your project consultation. Each group meets with me once, in weeks 4 through 6, for feedback on its plan: youcanbook.me.

Objectives#

  • Publish a docs/ folder as a live website: Publishing to GitHub Pages.

  • Write and compile LaTeX, and automate the inclusion of a figure or table so a PDF rebuilds itself from the pipeline’s outputs: Introduction to LaTeX and LaTeX Essentials.

  • Use the pull-request workflow: branch, PR, review, revise, merge (GitHub Issues and Pull Requests).

  • Explain how the Nelson-Siegel-Svensson model fits a yield curve to Treasury quotes, and how fed funds futures prices imply an expected policy path.

Agenda Item 1: Publishing to GitHub Pages#

Publishing to GitHub Pages: how a docs/ folder of static HTML becomes a live website (Settings → Pages → Deploy from a branch → main / /docs), and why your assignment repo must stay private while the site is public. The two steps compose: chartbook build produces docs/, Pages serves it. HW 2 Part 3 is exactly this, into a separate public repo.

Agenda Item 2: LaTeX for PDF Reports#

ChartBook is the HTML half of report generation; LaTeX is the PDF half.

  • Introduction to LaTeX and LaTeX Essentials. Compile the report templates from the project template, in the project you scaffolded last week with chartbook init.

  • Work through the latex/ examples in the in-class repo, a progression from a five-line document through figures and tables, bibliographies, and Beamer slides.

  • Demo: the market brief in latex/08_handout, built from a custom class file, whose charts and returns table are generated by a Python script and pulled in with \includegraphics and \input. The PDF rebuilds itself whenever the data changes. This is the same pattern the class report uses in week 6. → HW 2: the same pattern on the yield curve. task_generate_market_brief writes the charts and tables, and task_compile_market_brief runs latexmk on reports/market_brief.tex.

Agenda Item 3: Issues and Pull Requests#

GitHub Issues and Pull Requests: how contribution works in the open-source world, and the workflow you will use for the collaborative report and for your final project.

  • Issues as the coordination layer: claiming work, scoping it, linking fixes with fixes #123.

  • The PR loop, demoed live: branch → commit → open a pull request → review comments → revise → merge.

  • Why maintainers protect main and require review, and how these mechanics let strangers collaborate safely on the packages we depend on. This is the “peer code review” that shows up in the job postings.

Agenda Item 4: Launch HW 2, the Yield Curve and the Policy Path#

HW 2 is a multi-source pipeline that ends in a published website and a PDF. You estimate the U.S. Treasury yield curve with the Nelson-Siegel-Svensson model, following Gürkaynak, Sack, and Wright (2006), the methodology behind the Fed’s daily published curve, using the finm package for the fixed-income functions:

Then you read the expected path of the policy rate out of 30-Day Fed Funds futures, the way the CME FedWatch tool does:

The pipeline pulls from four sources (CRSP via WRDS, the Fed’s published GSW series, ZQ futures from Databento, and FRED), and the deliverables are a ChartBook site on GitHub Pages and a one-page market brief on the Treasury curve, compiled from LaTeX. We launch it in class; you finish the TODOs and publish on your own.

Final Project: The Two Steps#

The full procedure and rubric are in the Final Project Instructions and Rubric; we walk it in class. Two steps:

  1. Consultation. One-on-one with me, booked via youcanbook.me, in weeks 4 through 6. Come ready to say what the paper is about, your data sources, and the product you will build: the new table or figure, and who would use it.

  2. Final presentation and oral defense, week 10: the completed project, with each member individually quizzed.

Looking ahead to Week 4#

Week 4 turns the code itself into something installable and documented: Python packaging and Sphinx. HW 3 launches, an installable package that recovers option-implied crash probabilities from the volatility smile.

Contents