From CS 294-10 Visualization Sp10

Jump to: navigation, search


For this assignment, three key decisions were made in designing this image. The first was to use dot charts to represent the data. The second was to use paneling instead of representing all of the data on one graph. Finally, the last decision was to sort the countries in order of donation in each graph. We discuss the motivations for each of these in turn and conclude with the “story” that the graph is trying to tell.

As discussed in the reading, dot charts are an effective data representation mechanism for multiple reasons. In these charts, data is represented as position, which was considered in the reading to be the most effective encoding mechanism. Additionally, these charts promote high data-ink ratios (the ratio of the ink in the graph critical to data representation to the total) and low lie factors (i.e. changes in the graph accurately represent changes in the data), both of which are heavily advocated for by Tufte as necessary for the effective communication of values.

The data ink ratio could be increased further by removing the dotted lines associating the values with the countries. However, since this association aids the eye in understanding the graph, we kept it in. But, we did make the lines significantly thinner so as to not overwhelm the rest of the figure.

The second decision we made was to use paneling as opposed to presenting all of the data on a single graph. We made this choice because the per capita donations are miniscule compared to the country-wide donations. As such, we would have had to use two x-axis on the same graph or some other type of representation. We believe that this would have cluttered the graph and made it much more challenging to understand the data.

The third decision we made was to order the countries separately in each panel. Maintaining the order across panels would have made it easier for one to determine all of the statistics about a specific country. However, sorting each panel separately promotes an additional comparison across charts (i.e. the United States has donated the highest amount overall but it is middle-of-the-pack in terms of per-capita donation). We believe this additional level of comparison to be useful. Additionally, if one wishes to get all of the statistics about a country, a table is probably the best mechanism.

We conclude with a description of the data we chose to focus on and its impact on the graphs. First of all, our graphs only portray donations from countries (not donations from private entities, the world bank, ext). Focusing on these values avoids cluttering the image, whose intention is to compare country donations. At the same time, we do not believe that the remaining values would provide significant additional context to the reader. Secondly, our image focuses on the per-capita donations instead of the country population as we feel the former is more relevant in this context. Lastly, our graphs do not directly compare a country’s per-capita donation to its total donation. While we thought that this might be an interesting comparison, when we plotted the graph, we did not find an interesting trend. As such, we determined that it would not be useful to include this in the figure.

[add comment]
Personal tools