Lars Schöbitz - website & CV

A data-driven personal website and CV. Content lives in versioned CSV files under data/; a multi-page Quarto website (*.qmd) renders to docs/ for GitHub Pages, and a separate Typst document renders the PDF CV.

This replaces the earlier single-page CV (and, before that, the datadrivencv / pagedown Google Sheet setup). Old files are preserved under archive/.

How it works

data/*.csv  ->  R/cv_functions.R  ->  *.qmd  ->  docs/ (website) + cv.pdf

Every website page is a thin .qmd that sources the R helpers, calls create_cv("data") once, and prints only the sections relevant to that page. The section column in data/entries.csv routes each row to a page.

Pages

Page	Content
index.qmd	Start page (page-layout: custom): hero with portrait, role line, intro and contact links; live numbers; upcoming talks and workshops with full dates, event sites and materials links; recent talks; latest writing; explore cards; contact band. All from the aboutme package through R/cv_functions.R.
research.qmd	Publications, conference papers, technical reports, research interests, field research
software.qmd	Open data & software, grants, awards
teaching.qmd	Courses, workshops, talks
media.qmd	Press and video coverage (from data/media.csv)
blog.qmd	Blog posts as external links
cv.qmd	HTML CV summary with a PDF download button
cv-pdf.qmd	The full Typst CV; renders to cv.pdf. Not part of the website render list.

Data files

• data/entries.csv - every job, project, publication, report, talk. Columns include section (routes to a page), type (article, conference, report, …) and status (blank = published; under review / work in progress render as tags).
• data/media.csv - press/video coverage (outlet, date, kind, title, url).
• data/contact_info.csv, data/language_skills.csv, data/technical_skills.csv, data/text_blocks.csv - supporting data. text_blocks.csv holds the intro bio and the per-page intro and prose blocks (research interests, field research, etc.).
• data/braindump_entries.csv - work-in-progress rows you fill in by hand; rows still containing TODO are skipped automatically.

Assets

• assets/styles.css - site chrome and list patterns (navbar, pub-list, pub-tag, media-list, teaching-table, section-intro, cv-download-btn, skill bars).
• assets/theme.scss - maps the bitz brand palette onto the custom properties that styles.css reads, and keeps the web base font size at 1rem.
• assets/landing.scss - the start page classes (lse- prefix: hero, numbers, cards, rows, tint and band), modelled on the openwashdata and FAIRqual start pages.
• onepager/_brand.yml - the bitz brand file (colours, IBM Plex fonts). The website reads it through brand: in _quarto.yml, the one-pager's Typst template reads it directly. Keep it equal to bitzgmbh/brand (diff against a clone of that repo).
• assets/typst-helpers.typ - Typst template for the full CV PDF (cv-pdf.qmd) on the bitz brand: name block, uppercase section headings, dated entry grid, skill bars, footer. It mirrors onepager/onepager.typ, so both PDFs look alike.
• assets/profile-lars-schoebitz.jpg - profile photo (About page and favicon).

Building

./render.sh

This runs three steps: it renders cv-pdf.qmd to cv.pdf at the repo root, renders onepager/cv-onepager.qmd to onepager/cv-onepager.pdf (the one-page CV and publication list), then renders the website into docs/ (which picks up both PDFs as declared resources so the two download buttons on the CV page resolve). The root cv.pdf and onepager/cv-onepager.pdf are gitignored build intermediates; docs/cv.pdf and docs/onepager/cv-onepager.pdf are the tracked copies. docs/ is committed and is what Netlify serves.

To refresh harvested source data (optional): Rscript scripts/harvest.R re-pulls ORCID works and GitHub org repos into data/staging/ for review. Curated content stays in data/*.csv.

Deploy

• Netlify serves the pre-built docs/ folder from main (netlify.toml: publish docs, no build command). Merging a PR into main triggers a redeploy.
• Custom domain: www.lse.de (primary) and lse.de, configured in Netlify's domain settings; DNS A records at hosttech point both at Netlify.

Data sources

• Publications, reports, software, datasets: ORCID 0000-0003-2196-5015, cross-checked against the Zotero export data/my-articles.bib.
• Open-data projects & packages: GitHub orgs (openwashdata, rstatsZH, Global-Health-Engineering, ethopen).
• Roles, courses, workshops, education, media: hand-entered (data/BRAINDUMP.md).