Learning Tableau 10: Business Intelligence and data visualization that brings your business into focus, Second Edition

Tableau connects to data stored in a wide variety of files and databases. This includes flat files, such as Excel and text files; relational databases, such as SQL Server and Oracle; cloud-based data sources, such as Google Analytics and Amazon Redshift; and OLAP data sources, such as Microsoft Analysis Services. With very few exceptions, the process of building visualizations and performing analysis will be the same no matter what data source you use. We'll cover the details of connecting to different data sources in Chapter 2, Working with Data in Tableau.

For now, we'll connect to a text file, specifically, a comma-separated values file (.csv). The data itself is a variation of the sample data provided with Tableau for Superstore, a fictional retail chain that sells various products to customers across the United States. It's preferable to use the supplied data file instead of the Tableau sample data as the variations will lead to differences in visualizations.

The Chapter 1 workbook, included with the code files bundle, already have connections to the file; however, for this example, we'll walk through the steps of creating a connection in a new workbook:

We'll refer to elements of the interface throughout the book using specific terminology, so take a moment to get familiar with the terms used for various components numbered in the preceding image:

Now that you have worked through connecting to the data, we'll explore some examples that lay the foundation for data visualization and then move into building some foundational visualization types. To prepare for this, do the following:

With a connection to the data, you are now ready to visualize and analyze the data. As you start doing so, you will take on the role of an analyst at the retail chain. You'll ask questions of the data, build visualizations to answer those questions, and ultimately design a dashboard to share the results. Let's start by laying down some foundations to understand how Tableau visualizes data.

When you first connect to a data source, such as the Superstore file, Tableau will display the data connection and the fields in the data pane on the left sidebar. Fields can be dragged from the data pane onto the canvas area or onto various shelves, such as Rows, Columns, Color, or Size. We'll see that, placement of the fields will result in different encodings of the data, based on the type of field.

Measures and dimensions

The fields from the data source are visible in the data pane and are divided into measures and dimensions. The difference between measures and dimensions is a fundamental concept to understand when using Tableau:

• Measures: Measures are values that are aggregated. That is, they can be summed, averaged, and counted, or have a minimum or maximum.

• Dimensions: Dimensions are values that determine the level of detail at which measures are aggregated. You can think of them as slicing the measures or creating groups into which the measures fit. The combination of dimensions used in the view defines the view's basic level of detail.

As an example (which you can view in the Chapter 01 Starter workbook on the Measures and Dimensions sheet), consider a view created using the fields Region and Sales from the Superstore connection, as shown here:

The Sales field is used as a measure in this view. Specifically, it is being aggregated as a sum. When you use a field as a measure in the view, the type aggregation (such as SUM, MIN, MAX, AVG) will be shown on the active field. In the preceding example, the active field on Rows clearly indicates the sum aggregation of Sales: SUM(Sales).

The Region field is a dimension with one of four values for each record of data: Central, East, South, or West. When the field is used as a dimension in the view, it slices the measure. So instead of an overall sum of sales, the preceding view shows the sum of sales for each region.

Discrete and continuous

Another important distinction to make with fields is whether a field is being used as discrete or continuous. Whether a field is discrete or continuous, determines how Tableau visualizes it based on where it is used in the view. Tableau will give you a visual indication of the default for a field (the color of the icon in the data pane) and how it is being used in the view (the color of the active field on a shelf). Discrete fields, such as Region in the previous example, are blue, and continuous fields, such as Sales, are green.

Tip

In the screenshots, in the print version of this book, you should be able to distinguish a slight difference in shading between discrete (green) and continuous (blue) fields, but pay special attention to the interface as you follow along using Tableau. You may also wish to download the color image pack from Packt Publishing. You can click on the link: https://www.packtpub.com/sites/default/files/downloads/LearningTableau10_ColorImages.pdf.

Discrete fields

Discrete (blue) fields have values that are shown as distinct and separate from each other. Discrete values can be reordered and still make sense.

When a discrete field is used on the Rows or Columns shelves, the field defines headers. Here the discrete field Region defines column headers:

Here, it defines row headers:

When used for color, a discrete field defines a discrete color palette in which each color aligns with a distinct value of the field:

Continuous fields

Continuous (green) fields have values that flow from first to last. Numeric and date fields are often used as continuous fields in the view. The values of these fields have an order, which would make little sense to change.

When used on Rows or Columns, a continuous field defines an axis:

When used for color, a continuous field defines a gradient:

It is very important to note that continuous and discrete are different concepts from measure and dimension. While most dimensions are discrete by default and most measures are continuous by default, it is possible to use any measure as a discrete field and some dimensions as continuous fields.

Tip

To change the default of a field, right-click on the field in the data pane and select Convert to Discrete or Convert to Continuous. To change how a field is used in the view, right-click on the field in the view and select it to be either discrete or continuous.

In general, you can think of whether a field is continuous or discrete, as telling Tableau, how to display the data (header or axis, single colors or gradient) and measure or dimension, and how to organize the data (aggregate it or slice/group it).

As you work through the examples in this chapter, pay attention to the fields you are using to create the visualizations, whether they are dimensions or measures, and whether they are discrete or continuous. Experiment with changing fields in the view from continuous to discrete and vice versa to gain an understanding of the difference in the visualization.

A new connection to a data source is an invitation to explore. At times you may come to the data with very well defined questions and a strong sense of what you expect to find. Other times, you will come to the data with general questions and very little idea of what you will find. The data visualization capabilities of Tableau empower you to rapidly and iteratively explore the data, ask new questions, and make new discoveries.

The following visualization examples cover a few of the foundational visualization types. As you work through the examples, keep in mind that the goal is not simply to learn how to create a specific chart. Rather, the examples are designed to help you think through the process of asking questions of the data and getting answers through iterations of visualization. Tableau is designed to make that process intuitive, rapid, and transparent. Far more important than memorizing steps to create a bar chart is understanding how and why to use a Tableau to create a bar chart and then adjust your visualization to gain new insights as you ask new questions.

Bar charts visually represent data in a way that makes comparisons of a value across different categories easy. Length of the bar is the primary means by which you will visually understand the data. You may also incorporate color, size, stacking, and order to communicate additional attributes and values.

Creating bar charts in Tableau is quite easy. Simply drag and drop the measure you want to see on either the Rows or Columns shelf and the dimension that defines the categories onto the opposing Rows or Columns shelf.

As an analyst for the Superstore, you are ready to begin a discovery process focused on sales (especially the dollar value of sales). As you follow the examples, work your way through the sheets in the Chapter 01 Starter.twbx workbook. The Chapter 01 Complete.twbx workbook will contain the complete example, so you can compare your results at any time:

You now have a horizontal bar chart. This makes the comparison of sales between the departments easy. Notice how the mark type in the drop-down menu on the Marks card is set to Automatic and shows an indication that Tableau has determined that bars are the best visualization given the fields you have placed in the view. As a discrete dimension, the Department field defines row headers for each department in the data. As a continuous measure, the Sales field is defining an axis with the length of the bar extending from 0 to the value of the sum of sales for each department.

Line charts connect the related marks in a visualization to show movement or relationship between connected marks. The position of the marks and the lines that connect them are the primary means of communicating the data. Additionally, you can use size and color to visually communicate additional information.

The most common kind of line chart is a time series chart. Time series show the movement of values over time. They are very easy to create in Tableau and require only a date and a measure.

Continue your analysis of Superstore sales using the Chapter 01 Starter workbook that you saved earlier. The following are the steps to get the output of the Sales over time graph:

Observe the cyclical pattern that is quite evident when looking at the sales by quarter:

Tableau makes creating geographic visualizations very easy. The built-in geographic database recognizes geographic roles for fields, such as country, state, city, or zip code. Even if your data does not contain latitude and longitude values, you can simply use geographic fields to plot locations on a map. If your data contains latitude and longitude fields, you may use those instead of the generated values.

The power and flexibility of Tableau's geographic capabilities, as well as the options for customization, will be covered in more detail in the Mapping techniques section of Chapter 10, Advanced Visualizations, Techniques, Tips, and Tricks. In the following examples, we'll consider some of the foundational concepts of geographic visualizing.

Geographic visualization is incredibly valuable when you need to understand where things happen and if there are any spatial relationships within the data. Tableau offers two basic forms of geographic visualization:

Symbol maps

Another standard type of geographic visualization available in Tableau is a symbol map. Marks on this map are not drawn as filled regions; rather, marks are shapes or symbols placed at specific geographic locations. Size, color, and shape may also be used to encode additional dimensions and measures.

Continue your analysis of Superstore sales, following these steps:

1. Navigate to the Sales by Postal Code sheet.

2. Double-click on Postal Code under Dimensions.Tableau automatically adds postal code to the Detail of the Marks card and Longitude (generated) and Latitude (generated) to Columns and Rows. The mark type is set to a circle by default and a single circle is drawn for each postal code at the correct latitude and longitude. You may also notice an indicator for 1 unknown postal code. We'll take a look at how to handle this in the future chapters.

3. Drag Sales from Measures to the Size shelf on the Marks card. This causes each circle to be sized according to the sum of sales for that postal code.

4. Drag Profit from Measures to the Color shelf on the Marks card. This encodes the mark color to correspond to the sum of profit. You can now see the geographic location of profit and sales at the same time. This is useful because you will see some locations with high sales and low profit that may require some action.

The final view should look like this after making some slight adjustments to size and color details:

Sometimes you'll want to adjust the marks on a symbol map to make them more visible. Some of the options are:

• If marks are overlapping, click on the Color shelf and set transparency to somewhere between 50% and 75%. Additionally, add a dark border. This makes the marks stand out and you can often discern any overlapping marks much better.

• If marks are too small, click on the Size shelf and adjust the slider. You can also double-click on the Size legend and edit the details of how Tableau assigns size.

• If marks are too faint, double-click the Color legend and edit the details of how Tableau assigns color. This is especially useful when you are using a continuous field that defines a color gradient.

A combination of tweaking the size and using Stepped Color and Use Full Color Range, as shown here, produced the final result for this example:

Unlike filled maps, symbol maps allow you to use size to visually encode aspects of the data. Symbol maps also allow for greater precision. In fact, if you have latitude and longitude in your data, you can very precisely plot marks at a street address level of detail. This type of visualization allows you to map locations that do not have clearly defined boundaries. Notice that if you were to change the mark type from Automatic to Filled Map in the preceding view, you would get an error message indicating that filled maps are not supported at the level of detail in the view.

Show Me is a powerful component of Tableau, which arranges selected and active fields into the arrangement required for the selected visualization type. The Show Me toolbar displays small thumbnail images of different types of visualizations, thus allowing you to create visualizations with a single-click. Based on the fields you select in the data pane and the fields that are already in the view, Show Me will enable possible visualizations and highlight a recommended visualization. Explore the features of Show Me by following these steps:

Observe that the Show Me window has enabled certain visualization types, such as text tables, heat maps, symbol maps, filled maps, and bar charts. These are the visualizations that are possible given the fields already in the view, in addition to any selected in the data pane. Show Me highlights the recommended visualization for the selected fields and also gives a description of what fields are required as you hover over each visualization type. Symbol maps, for example, require one geographic dimension and up to two measures.

Other visualizations are grayed out, such as line charts, and histograms. Show Me will not create these visualization types with the fields that are currently in the view and selected in the data pane. Hover over the grayed out line charts in Show Me. Show Me indicates that line charts require one or more measures, which you have selected, but also require a date field, which you have not selected.

Show Me can be a powerful way to quickly iterate through different visualization types as you search for insights into the data. But as a data explorer, analyst, and storyteller you should consider Show Me as a helpful guide that is giving suggestions. You may know that a certain visualization type will answer your questions more effectively than the suggestions of Show Me. You may also have a plan for a visualization type that will work well as part of a dashboard but isn't even included in Show Me.

Conclude the Show Me example by experimenting with Show Me, clicking various visualization types and looking for insights into the data that may be more or less obvious based on the visualization type. Circle views and box and whisker plots show the distribution of postal codes for each state. Bar charts easily expose several postal codes with negative profit.

Often you'll need more than a single visualization to communicate the full story of the data. In these cases, Tableau makes it very easy for you to use multiple visualizations together on a dashboard. In Tableau, a dashboard is a collection of views, filters, parameters, images, and other objects that work together to communicate a data story. Dashboards are often interactive and allow end users to explore different facets of the data.

Dashboards serve as a wide variety of purposes and can be tailored for a wide variety of audiences. Consider the following possible dashboards:

Considerations for different audiences and advanced techniques will be covered in detail in Chapter 7, Telling a Data Story with Dashboards. For now, follow these steps for an example that introduces the foundational concepts:

The dashboard window consists of several key components. Techniques for using these objects will be detailed in Chapter 7, Telling a Data Story with Dashboards. For now, focus on gaining some familiarity with the options that are available.

The left sidebar contains two tabs:

The Dashboard pane contains options for previewing based on target device, sizing options, and a list of all visible sheets (views) in the dashboard. You can add these sheets to a dashboard by dragging and dropping. As you drag the view, a light grey shading will indicate the location of the sheet in the dashboard once it is dropped. You can also double-click any sheet and it will be added automatically.

The next section lists multiple additional objects that can be added to the dashboard. Horizontal and Vertical layout containers will give you finer control over the layout; Text allows you to add text labels and titles. Images and even embedded web content can be added. Finally, a Blank object allows you to preserve blank space in a dashboard or can serve as a place holder.

Using the toggle, you can select whether new objects will be added as Tiled or Floating. Tiled objects will snap into a tiled layout next to other tiled objects or within layout containers. Floating objects will float on top of the dashboard in successive layers.

Tableau's visual environment allows for a rapid and iterative process of exploring and analyzing data visually. You've taken your first steps in understanding how to use the platform. You are connected to the data and then explore and analyze the data using some foundational visualization types such as bar charts, line charts, and geographic visualizations. Along the way you focused on learning the techniques and understanding key concepts, such as the difference between measures and dimensions, and discrete and continuous fields. Finally, you put all the pieces together to create a fully functional dashboard that allows an end-user to understand your analysis and make discoveries of their own.

In the next chapter, we'll explore how Tableau works with data. You will be exposed to the fundamental concepts and practical examples of how to connect to various data sources. Combined with the foundational concepts you just learned about building visualizations, you will be well equipped to move on to more advanced visualizations, deeper analysis, and telling fully interactive data stories.