Session 3B: Streaming Data D3

Session 3B: Streaming Data + D3

Notes

• Use websockets and add D3 update logic into onMessage
• Use the same method for updating from streaming or rest
• traditional way to update might be to replace an array, while with streaming, you likely only change some elements
• choosing a proper update interval is important and can be domain specific
• streaming data can give a sense of the 'velocity' of how things are changing in a visualization
• practical applications: managing high traffic in a facility, development/ops deployment feedback, groups of people who need to see something happening together
• overhead of garbage collection can be a problem if updates/mutations are frequent
• batching updates needs to be explored for efficiency - especially for server data sending load, summaries of data over time periods.
• update ticks and different browsers can often create unexpected bugs
• svg rendering can be expensive and sometimes not practical for many nodes

Thoughts

• You may be presenting data in real-time, but what is the possible reaction time for someone to consume that information and respond?

Resources

• ImmutableJS https://facebook.github.io/immutable-js/ can be useful, but overhead can be costly.
• D3 Bubble Chart with PubNub Live JSON Demo

Examples

• Fun with D3.js: Data Visualization Eye Candy with Streaming JSON[]
• https://www.instagram.com/p/zGUzZcKvUu/

No link demos

• conversations on twitter, people who are talking about a topic and how they are related
• realtime views of customer usage of an app