University of Southampton OCS (beta), CAA 2012

Font Size: 
Visualising time with multiple granularities: a generic framework.
Jean-Yves BLAISE, Iwona DUDEK

Last modified: 2011-12-21


Investigating the evolution of historic architecture or archaeological sites often starts with the cumbersome task of putting together various pieces of information, each with its specific precision, scope and reliability. Naturally, time slots are among the main clues analysts expect to spot when filtering and cross-examining.

But due to the very nature of historic data sets – heterogeneity, uncertainty, missing data (etc.) – time points and intervals the analyst will identify are often inconsistent in terms of granularity. Time intervals, typically the overall lifetime of an artefact, its periods of construction or modification; may be described by expressions like “between the last quarter of the XIIIth c. and the middle of the XIVth c.”, whereas some punctual events may be recorded more precisely, in cases like “fire on Nov 29th 1554” or “town’s siege between march 1445 and November 1445”.

In parallel, describing an artefact’s life often implies taking into consideration pieces of information that correspond to regular or cyclic events, where there are also inconsistent granularities. Typically, when analysing an isolated chapel at high altitude, the analyst copes with a fuzzy cyclic behaviour – the chapel is inaccessible due to snow for a certain number of weeks during the year – as well as with a well-defined cyclic behaviour – a pilgrimage is organised yearly. 

In other words, may it be because of the nature of historic data sets, or may it be because of the heterogeneity of the events we need to report there are very few solutions analysts can count on if they need to visualise in a consistent, insight-gaining manner the time slots they have spotted?

In historic sciences, it appears clearly that the handling of multiple time granularities is one of the major bottlenecks in the analyst’s visualisation effort. Our research aims at giving analysts means to combine in a single visualisation multiple aspects of the parameter time, and particularly multiple granularities.

As a first step, we initially focused on a visual comparison of 25 alternative calendars covering a wide range of historic periods and civilisations (Julian and Gregorian calendars of course , but also Babylon, ancient Egypt, Japan, Incas, Inuits, Burma, etc.). The visualisation sums up thirteen key aspects of calendars (cycles, divisions, period and area of validity, correction mechanisms, etc.) corresponding to alternative time granularities. The visualisation helps underlining legacies, alternative visions of time as linear or cyclic, common or opposing choices, alternative divisions of the year, etc. A prominent service offered by this visualisation is that it enables comparisons at various time granularities “within the eyespan”, to quote E.R Tufte.

So this first result has been extended to propose a more generic framework for visualising time with multiple granularities. It is applied on two very different test cases – the analysis of an individual artefact’s lifetime, and a comparison of stylistic trends inside / between territories.

The contribution will present the concepts and ideas behind this research, as well as their practical applications on the tests cases and accordingly its possible benefits for researchers and practitioners in historic sciences.


time granularity information visualisation visual reasoning