time-i-gram: A Grammar for Interactive Visualization of Time-based Data

The visual analysis of time-oriented data demands capabilities that general-purpose visualization grammars do not provide: multi-scale encoding that spans orders of temporal magnitude, native support for cyclic layouts that reveal periodic structure, and coordinated interaction that links heterogeneous variables at a shared temporal resolution. We present time-i-gram, a declarative visualization grammar that addresses these gaps through six components designed around the semantic structure of time: data types and time primitives, visual marks, visual channels, scales, layouts, and interaction. A prototype implementation extends Gosling.js, inheriting a scalable tile-based rendering engine originally developed for genomics. We demonstrate the grammar's expressiveness through three usage scenarios in epidemiology, renewable energy monitoring, and transportation-producing visualizations that are either unavailable or prohibitively complex in current grammars.