CS 767/867 Assignment 3: Due 7 October
A univariate map is a map for which a single value is shown at every point over a plane (also called a scalar field). A Bivariate map is a map that has two values displayed at every point on a plane. For example, we might wish to show both atmospheric pressure and temperature on a weather map. Or we might wish to show both climate zone and type of agriculture on a map of the US. The need to display two and more variables over a plane is very common in Geographic Information Systems (GIS).
In this assignment you will design and implement a weaving scheme to display a bivariate map and conduct an evaluation study. You should modify the code provided in two ways.
1) Design and implementation of a pair of color sequences that show as many distinct levels of the two variables as possible, providing the finest spatial resolution, and maintaining monotonic perceptual ordering if possible.
2) The design and implementaiton of code to run an experiment. This experiment will have three conditions.
Cond 1: Blended to red and green (already implemented).
Cond 2: The weaving color scheme already provided.
Cond 3: A new weaving color scheme of your own design.
The code you impement to run the experiment should take the subject through a set of 120 trials in each condition, given in blocks of 40. A new map should be generated every 10 trials. The software will record the results to a file for later analysis.
On each trial, the software positions a cursor over part of the bivariate map, the subject clicks a corresponding part of the Key.
Your report will have the form of a standard scientific report of an experiment. It will have the following sections and amount to < 1200 wds.
Title:
Introduction: The introduction should give a brief intro to the problem citing at least one relevant research paper. You should introduce your texture coding scheme here.
Method: This should describe the critical aspects of the experiment.
Results: Give at a minimum mean absolute errors and the error variancefor each condition - an Excel spread sheet is a good way to do the analysis. Note: there are three different error values that can be given for each condition (as per class discussion).
Discussion: You should discuss the significance of the results and perhaps suggest improvements to your study.
References:
Miller, J.R. (2007) Attribute Blocks: Visualizing Multiple Continuously defined Attributes. IEEE Computer Graphics and Applications. May-June, 57-69.
Hagh-Shenas, S. et al (2007). Weaving versus blending: a quantitative assessment of the information carrying capacities of two alternative methods for conveying multivariate data with color. v13(6) 1270-1277.