DataScience.LA archive

A read-only archive of DataScience.LA, the blog of the Los Angeles data science community (2014–2021): meetup talks, useR! 2014 interviews, benchmarks and essays.

The site is a static Hugo site using the Blowfish theme, deployed to GitHub Pages by GitHub Actions on every push to master.

Layout

Path	What
content/posts/<slug>/	One page bundle per post: index.md, its images, and comments.json (archived comments)
content/team/, content/join-membership/	Pages
content/authors/, content/categories/, content/tags/	Term pages (names, slugs, WordPress URL aliases)
data/authors/	Author names, bios and photos used by Blowfish
layouts/	Site templates: archived comments, link handling, embed shortcodes
assets/	Logo, author photos, CSS overrides (assets/css/custom.css)
config/_default/hugo.toml	Site configuration
themes/blowfish	Theme (git submodule, pinned)
tools/wordpress-import/	Scripts that generated content/ from the WordPress database dump

URLs match the original WordPress site: posts at /<slug>/, categories at /category/<slug>/, tags at /tag/<slug>/; author pages are at /authors/<slug>/ with redirects from /author/<slug>/.

Building locally

git clone --recurse-submodules git@github.com:DataScienceLA/datasciencela.github.io.git
cd datasciencela.github.io
hugo server            # http://localhost:1313/ — needs Hugo extended ≥ 0.157
hugo build --gc --minify   # output in public/

Shortcodes

Shortcode	Use
{{< youtube-lite id="…" title="…" >}}	YouTube video, loaded (from youtube-nocookie.com) only when clicked
{{< speakerdeck id="…" url="…" title="…" >}}	Speaker Deck slides
{{< tweet-archive url="…" author="…" date="…" >}}text{{< /tweet-archive >}}	A tweet, preserved as text
{{< iframe src="…" >}}	Other embeds (maps)
{{< avatar src="…" >}}	Round photo on the Team page

How the content was produced

content/ was generated from the WordPress database dump of the original site (not included in this repository) with the scripts in tools/wordpress-import/:

1. extract.py reads the SQL dump and writes published posts, pages, authors, categories, tags and approved comments to JSON (no e-mail addresses or other private data).
2. convert.py turns each post into a Hugo page bundle: WordPress paragraph formatting, shortcodes (captions, Symple columns/headings, Easy Table, Speaker Deck, gists), auto-embedded YouTube/Speaker Deck/Twitter URLs, images (full-size originals, resized to ≤1600 px) and comments.

Embedded GitHub gists were copied into the posts as code blocks. Tweets were preserved from WordPress' oEmbed cache. Posts can be edited directly as Markdown now; the scripts are kept for reference.

Videos and slides that were no longer available when the archive was made (shown as a note with the original link): YouTube AQ0EjOJnu9Q, 5_vpPJPA9UQ, a6VRHvSapYQ, 8x3nVP9B0hQ, DZ0Y7fbipXs, 6DcUNKBdHiQ (deleted), H_YtvFxD8ig, d0aVup0QuuI, lZ6q1Orwtug (private), and the Speaker Deck deck szilard/machine-learning-meetup-febr-2017.