Work on TLCMap will commence in 2019 and continue into 2020.
Various tools will enable visualisation of quantitative data in different ways.
Developed for science, engineering and commercial purposes, many mapping system already have extensive visualisation features for quantitative data. Rather than attempt to replicate all this functionality, we aim to provide some simple tools to:
Some specific requirements we aim to address include:
A common problem in humanities is dealing with vague and uncertain data. Narratives may contain comments like, "It happened in late Winter, a day's ride north of the creek." A manuscript might be dated to sometime in the life of a medieval poet, and there may be some debate over their birth and death date. They might have written it in one of 2 cities. We might want to indicate that there are no distinct boundaries between languages but mapping systems constrain often us to draw clear and distinct regions. And so on. Yet computers, to work with dates and places, need specifics. Some times we can work around this by specifying ranges, such as translating 'late March, 1836' to a range of between 15/03/1836 and 31/03/1836, or by specifying our best guess and adding a 'notes' field with commentary on the accuracy. We hope to ease this situation by finding ways and practices for representing such vagueness - perhaps with blurs, colour coding, thickness and so on.
Often we make maps to see or show patterns. Statistical comparison can give empirical weight to these patterns, reveal patterns we hadn't noticed and provide better quantitative comparisons. They can help with common problems in maps - for example if we wish to base an argument on the fact that some set of events occurs close in time and space to another set of events, from which we might infer or make a case for some causal connection. Statistics can help us measure how 'close' two sets of data are, and in relation to other sets of data. When dealing with maps we are typically dealing with spatial coordinates, often scattered sets of points, and so the full range of statistical analysis should be open to us. However, basic statistics is typically done on 2D surface infinite in all directions whereas geographical data is on a 3D ellipsoid surface. The near the international date line, the longitude -179 and +179 are only 2 degrees apart (very close) not 358 (very far) so even the most basic statistical calculations need to be adjusted. We hope to provide some basic tools for common statistical measures using coordinates.
Some assertions we'd probably want to be measurable include:
Often we make maps to see or show patterns. But when we see a cluster of points are they close together, or is it just travel time that makes them look that way, relative to some other points. One day's travel in mountains cover's less space than 1 days travel on the plains, so places that look close together in the mountains may really be the same distance, factoring in travel time, as those that look far apart on the plains. There may be other factors that affect this, or we might want to use gradients of terrain as a rough indication of travel time. Whatever it is we should be able to warp space according to time, or some other factor to get a better understanding of the patterns.
A technique emerging from archaeology and 'wayfinding', analyses terrain to find routes through rugged terrain and around obstacles by calculating trade offs of slope and distance to show likely routes. This assumes for example, that a walker would choose to save energy by talking the 'path of least resistance', ie: flatter and the least distance. Often this technique assumes valley walking but walkers, at least in Australia, often take the ridge route because there's less obstruction than gullies and you can see the distance better, and the ridges aren't often snow covered cliffs. So we'd like to automate wayfinding given terrain data for ridge and valley routes.
Time Layered Cultural Map is funded by the Australian Research Council, PROJECT ID: LE190100019
