Getting Into Tableau: Visualizing Iris Data Set

Introduction

Tableau has a tagline which goes ‘Changing the way you think about data.’ And on some contemplation, it can easily be concluded that that tagline is so much on point. Tableau as a tool has given us the power and shown us the mysteries that a data holds. More importantly, Tableau has taught us how to harness that power of data and uncover those mysteries. So, In this article Getting Into Tableau, I will take you through the important features of Tableau using Iris Dataset.

Now Tableau is much more than a data visualization software, it is also a business intelligence application (BI App). With Tableau, business intelligence becomes a self-service.

Essentially, Tableau is a data visualization software which is focused on business intelligence. Now, for some of you those two words, ‘business intelligence’ might seem to be confusing and they might even overwhelm a few. But the idea of simply supporting or helping an organization make its decisions, be it anything from operational to strategic, with the help of data visualization i.e. presenting data in a visual form, is what business intelligence is at its core.

Tableau is a clear leader in empowering the entire enterprise of business intelligence with modern analytics. According to the Gartner’s Magic Quadrant 2017, Tableau is a clear leader in business intelligence and analytics platforms.

Now that we have this idea as to what Tableau is, let us now understand its user interface to finally be able to use the software.

Tableau’s Interface

The anatomy of Tableau’s workspace is pretty easy to understand and it is made with a simplistic approach, even then it fails to be as simple as simple can be and it may take some getting used to.

Let us just dive into it.

At the top of the page, we have some dropdowns for the file, data, worksheet, dashboard, et cetera. And with regular use, you will realize that the most valuable commands exist in these drop downs.

Underneath these drop-down menus, we have a toolbar and the toolbar includes a number of useful commands. The first of these is save, add a new data source, pause any auto updates you have from your data sources, add a new sheet, duplicate and a couple of other commands that will come into play when we’ve created a visualization.

The Tableau logo command, on the left corner, brings you to the start page which you might need to return to if you want to connect to another data source.

To return, we simply hit the logo again.

Further, we have the data pane on the left-hand side of the page. If we are on the data tab, we’ll see that it lists all the open data sources and selects the fields from that data source as either dimensions or measures.

Measures are continuous variables and, essentially, numbers. Whereas, dimensions are discrete variables such as text strings and dates.

The data pane will also show any sets or parameters we may have.

If we click on the analytics tab we can bring out pieces of our analysis directly as drag-and-drop elements. If they’re not relevant to the type of view or visual that we are currently working on, certain elements will be greyed out. For example, totals on a timeline.

If we select something like a trend line we can bring it out to any of these drop areas to control aspects of its properties like model type and which measure it should apply to etc.

At the bottom, we have the new sheet tabs. We can create sheets, dashboards, and stories with these tabs. We can do things like rename the sheets, duplicate sheets, copy formatting, and many other things.

If the workbook has a lot of sheets we can navigate easily with the controls in the bottom right corner.

One of the most important parts of the interface is the rows and columns shelf in the right-hand center of the interface. To create our visualizations, we simply drag our dimensions and measures from the data pane into the columns and rows shelf. We can also do this by dragging our dimensions and measures directly onto the canvas and let Tableau segregate the rows and the shelves for us automatically.

When we drag the dimensions and measures to the columns and rows card we will observe a particular color scheme. The blue color will be for the dimensions and green will be for the measures.

For almost every visualization you create we will add dimensions and measures to the columns shelf and the rows shelf.

This will be clarified more when we take an actual example and a see first-hand application.

Then we have the central work area where the visualizations will appear once there are measures and dimensions in the rows and the columns shelves.

Here we can give the sheet title, write captions about our visualization, arrange our visualizations and also do some formatting.

On the top right corner, we have the Show Me option.

If you’re not sure what visualization to create or what options are available in Tableau, you can click the Show Me option, which will then display various type of visualization you can create with your selected dimensions and measures or from the data currently housed in columns and rows.

One nice aspect of the Show Me option is that it tells you automatically which types of charts are appropriate for the data that you have. Hence, Tableau tries to direct you to some best practices in data visualization.

Now coming to one of the most important parts of the Tableau interface. The pages, filter and the Marks cards on the center left of the page, next to our rows and columns shelves and the work area.

Both of these cards, filter, and marks, are used extensively on most visualizations.

Filters allow us to exclude certain data types from our visualization. For example, we might want to exclude some values from our visualization and the data used for that particular visual.

To do that we can set a range and removes values.

Marks, on the other hand, allow us to format the chart in more detail. Marx card is made up of several other shelves each of which can have fields placed on them and can be clicked on to edit their characteristics. Changing the mark type can change the shelves on the marks card such a selecting shape brings up the shape shelf.

Depending on the composition of the view there can be multiple marks cards, one for each measure.

Legends such as for color size and shape will automatically be created when a field is placed on the color size or shape shelf. However, legends can be removed by clicking on the menu and selecting hide card. To bring a legend back right-click anywhere off the canvas itself select legend and choose the one you want.

Application: Creating Our First Visualization Using Tableau

To bring more clarity and develop an enhanced understanding of what we have just learned about the interface of Tableau, let us see an example and create our first visualization.

For this first visualization, we will create some simple charts using the very famous Iris Dataset from the UCI Machine Learning Repository.

https://archive.ics.uci.edu/ml/datasets/iris

This is a fairly simple dataset with 150 instances and 4 attributes and a class; namely, sepal length, sepal width, petal length, petal width and the class attribute which has three classes, namely, Iris Setosa, Iris versicolor and Iris Virginica.

Even though we won’t be carrying out any classification or predictive analysis with this dataset at the moment, but this dataset has been widely used to carry out the same.

Off the bat, just as we imported our excel sheet with the Iris dataset into Tableau, we can view the various instances and attributes of our dataset. Here, we can also change the data type of the various attributes.

Now let us go directly to the worksheet and start with our visualization.

For the first example here, we see a plot between the petal length and the petal width which is color-coded according to the species, which can be seen in the legend.

Here, in the data pane, we see the division of the attributes between measures and dimensions. As explained earlier, the numeric attributes went into the measures category and the class attributes went into the dimensions category.

We have created a very simple scatter plot which shows the relationship between the petal length and the petal width taking into account the various species. Which has been placed in the columns and rows shelves respectively?

In the marks card, we can clearly see that the specified attribute has been dropped over the color shelf to color code the scatter points according to the species. We can also see a complementary legend for the color codes.

—————————————————————————————————————————————————-

Let us attempt another visualization, this time, histograms.

This visualization represents a comparison between the three species on the basis of their petal length, petal width, sepal length, and their sepal width.

We can see that these four attributes are a part of the rows shelf and the class variable, species is a part of the Columns shelf.

Further, the color coding has been done according to the species by using the marks card, as done in the last example and we can also see the species on the X-axis of the graph as species is a part of the columns shelf as well.

Conclusion

Now that we have seen these two examples, the usability of Tableau and its interface must be pretty clear.

Tableau has a lot of functionalities and lots of small adjustments can be made by using various options which you will only come across and while practicing making your own visualizations.

There a lot that needs to explore, so hit up UCI Machine Learning Repository or any dataset source like Kaggle and take your first steps into this world of visualizations and be the explorer this field requires you to be.

You might also like More from author