[MUSIC] Hello and welcome back. In the last lesson, you were asked to consider the context and potential workflows for your audiences. In this lesson, I want to illustrate some considerations for selecting the appropriate charts and graphs for your visualizations. In a sense, the core of data visualization is revealing relationships among different dimensions in the data. These relationships each have corresponding data types and design options associated with them. For example, to show change over time, line charts are a simple and effective choice. However, to display relationships and categorical data, bar charts work well but line charts don't. Once you know the possibilities for chart types to show instead of relationships, you can then decide which one or even multiple appropriate types would be best for the specific audience. In this lesson we briefly touch in a few examples, let's get started. Welcome back, in the last lesson you were asked to consider the context and potential use cases of your visualization for your audiences. In this lesson, I want to illustrate some visualization approaches to display a few primary kinds of data types and relationships. Awareness of these options and how to select the appropriate chart types based on data type in a relationship, can be an essential part of your visualization design work. In a sense, the core of data visualization is about revealing relationships among different dimensions of data. These relationships all have corresponding design options associated with them. For example, line charts are good at showing change over time, whereas bar charts are meant for showing categorical comparisons. Different types of charts are good for showing different types of data and kinds of relationships, but not all. You should be aware of what options work well for a particular situation and just as importantly, what doesn't. Let's begin with a simple example to illustrate the point. If you want to visualize, say, the proportion of apples versus oranges sold at Francesca's Fruit Stand at a particular date. Can you think of a chart type that would work, and one that wouldn't in this scenario? Well, a bar chart is one good option, but a trend line wouldn't work at all, why? Well, the numbers of apples and oranges are discrete data, that is, they are very distinct and distinguishable items. It's either an apple or an orange, there's no range between the two, there's no spectrum. Another example is, if you flip a coin, you can count and compare the number of heads and the number of tails, but there's no intermediate heads tails range between the two possibilities. So when comparing apples to oranges, or heads to tails, think discrete, and think bars, among other options. Now let's take a look at another type of data called continuous, which is, in a sense, the opposite of discrete data. As the name suggests, for continuous data, there's a connected range of values. So, for example, there's a popular baked good called the Twinkie that has a notoriously long shelf life in the package. How long can a Twinkie be safely consumed and how can that be shown? That's a very indiscreet question, it can be measured and displayed in a line starting from the date of manufacture and continuing to some point in the distant future in this case. Again, a line is the more appropriate choice to show this sort of trajectory. In addition to thinking about data types, you need to consider the kinds of relationships you want to display, and the best chart choice to do, as we said. In many cases, you'll have more than one option to choose from, and that's great. Once you know the possibilities for chart types to show a set of relationships, you can decide which one or even multiple appropriate forms will be the best for the audience. There are various guidelines that can help you find the right chart for the situation but let's take a basic example. Here are a few primary types of relationships and associated charts. For comparison, column charts are good. For composition, or part-to-whole relationships there's stacked columns and tree maps. Correlations, there are scatter plots, trends, line charts, ranking, bar charts, distributions, histograms and box plots, geospatial of course there are maps. There are many, many chart types, the key is to really understand which ones will work for the situation. This is only scratching the surface. We'll dive deeper into matching and combining chart types and data relationships in Course 3. Visualization designers, you really need to start make decisions based on data type, the kinds of relationships, and the best option to use for your intended audience. Thanks very much, and see you again.
