Global Temperature Data Visualization · Hower Chen
Hand-drawn sketches exploring different visualization approaches for the global temperature dataset.
Scatterplot showing Land Average Temperature over time. Circle radius encodes the temperature uncertainty (LandAverageTemperatureUncertainty). A semi-transparent Earth image provides contextual background.
Dot Size: The radius of circles is mapped to the uncertainty in temperature estimates (LATU). This way, the viewer immediately gets an idea of the uncertainty of the results. The larger the circles, the higher the uncertainty, and the smaller circles mean lower uncertainty. As the dates gets more modern, smaller circles are drawn, showing an improvement in the coverage and methodologies used for estimation.
Color Encoding: The color gradient from blue to red is used to encode the temperature data. By default, blue tends to indicate lower values, and red typically indicates higher values, which makes the visualization easy to read. This is one form of redundancy, where we have replotted the y-axis values (temperature) in the form of a color map, allowing the plot reader to quickly know the trends without always going back to the y-axis.
Background Image: The background image is the globe with a thermometer which shows the audience what we are trying to show. The reason is that the chart represents temperature measurements all over the globe. However, placing this image might cause visual noise or reduce readability of the points on the chart.
Overall Design: From the visualization, it is seen that the visualization provides information about short-term oscillations as well as long-term trends over a period of time. When scanned from left to right, it can also show information about the rising temperature trends. Similarly, dot size shows information about changes in uncertainty over the course of history.