Django: a Data Shovel With a Future
summary
Claims that many web UIs become bottlenecked by server-side multiway joins (often 5–12-way) and an enormous URL state space (example calculated as 75 million states). Built a Django-backed site for the New England Wildflower Society where initial design did expensive server joins; pivoted to ship per-filter JSON lists to the client and use JavaScript/underscore.js to intersect them locally. Result: no costly search URL, trivial caching (CDN/varnish + cache-control), lower server load and a site simple enough to be served statically — a true data shovel approach.
AI-written summary · about
chapters (15)
- 0:00Intro
- 1:43Once my code is finally working
- 3:42Web framework history follows a similar pattern
- 8:42A web framework denormalizes data to documents
- 10:02How much data should you deliver to the client?
- 12:54if hard drives kept growing
- 17:05Subversion git/hg minimum maximum
- 20:09NEW ENGLAND WILD FLOWER SOCIETY
- 21:27Problem: dichotomous keys
- 26:56red flowers smooth leaves Vermont
- 42:48Be aware of Python's growing data toolset
- 46:42NumPy arrays Pandas dataframes
- 49:55Python is becoming an important tool for large data sets
- 50:43URLs are simply text code + data
- 51:39Views can be simple procedures
related talks
31mDjango for life (sciences)
54mDjangoCon 2008 Panel: Django in Journalism
23mThree Django Apps in a Trenchcoat: Writing an event management website with not a lot of time
26mBuilding Dynamic Dashboards With Django and D3
26mModern web apps with (almost) no JS: Enhancing UX with HTMX in Django
52mChallenges in developing a large Django site
21mBuilding maintainable Django projects: the difficult teenage years
26mDjango at