The Loader is a Lie: How to Elegantly Load a Million or More Data Points

on YouTube ↗47m

" summary

Claim: loaders lie but can be fixed by showing real progress and streaming results. The team queries a million+ points from BigQuery, pages results via the Node.js client, and streams them with RxJS into charts and a WebGL map layer using deck.gl and Mapbox. They split requests into chunks, render partial charts and skeletons, animate a true percentage/timer, and demonstrate an interactive million‑point map that mostly works despite live‑demo delays.

AI-written summary · about

" chapters (4)

1. 0:00Start
2. 1:31Introduction
3. 5:42Loading spinners today
4. 11:38Improving the loading spinner

" related talks

• 37mVisualizing Large Datasets with JavaScript Using deck.glMarko Letic · RVA JavaScript Conf 2019
• 40mVisualizing Large Datasets with JavaScript Using Deck.glMarko Letic · code.talks 2019
• 30mSeeing the needle AND the haystack: single-datapoint selection for billion-poin...Jean-Luc Stevens · PyCon DE & PyData Berlin 2022
• 49mSaving the world with JavaScript: A Data Visualization storyMarko Letic · JS Fest 2019 Autumn
• 46mExplorations in Interactive Visual AnalyticsDoug Talbott · GOTO Copenhagen 2014
• 34mData Loading for Data EngineersWeston Pace · PyData Seattle 2025
• 31mBuilding a User Maintainable Visualization Dashboard for...Dharhas Pothina · PyData Austin 2019
• 30mMaps Platform: A Deep Dive on Building for Performance and ScaleTravis McPhail · Google I/O 2019