Visualizations comprise a cornerstone of the data science process and community. While datasets must be scraped, cleaned, modeled, and analyzed, these projects will ultimately fail to draw many people in without the lure of attractive and engaging visualizations. Throughout my time in Flatiron School’s data science bootcamp, I have enjoyed experimenting with different types of charts, graphs, maps, and other types of visualizations, in efforts to represent a dataset in the most effective and engaging way. However, above all else, my favorite has become the heatmap, in it’s unique multi-dimensional ability to represent several perspectives at once, all in the same visualization. As I polish up my final capstone project, I have come to realize that I hold a soft spot for this method of data representation, and feel no project is complete without at least one heatmap to show for it.
By definition, a heatmap is “a graphical representation of data where values are depicted by color.” With the added dimension of color, data scientists can utilize heatmaps not only to represent the relationship between two variables, but also to feature the weight, significance, or degree of influence of a particular point via the shade of color attributed to that value. Thought to have originated sometime in the 18th or 19th century, early users of the heatmap added different shades of grayscale to their charts in order to represent population density, ocean tide levels, and financial information. Now, with the rise of the technological revolution, programs such as seaborn, plotly, matplotlib, and R allow us to very spefically assign shades based on a given colorscale to compare different values against each other. The key step in the data science process allows us to present our findings in the most engaging, accessible format for all audiences, and increases the influence and efficacy of our work.
Similar to the heatmap, the choropleth map adds a geographic dimension to our already multi-dimensional visualization. Especially throughout my exploration of data from the modern Olympic Games, I found this relative of the heatmap to be extraordinarily effective in analyzing the games as a culture landmark that crosses and is defined by our international relationships and boundaries. Below, I display and briefly explore different heatmaps that I have utilized through several projects during Flatiron School’s data science bootcamp program, as well as some intriguing versions of the heatmap from other data scientists.