What is Tableau in Data Science? Uses of Tableau Software Tool

The literal meaning of the word ‘Tableau’ is ‘picture’ or figure that represents a story or scene. That’s precisely what the software tool means, too, as we will explore in this article.

Apart from understanding what Tableau is, we will also explore the different flavors it comes in, why learning Tableau is essential both for programmers and non-programmers, and then dive into a detailed example (using the free version of the software) to show how easy it is to work with the popular data visualization and BI tool.

Tableau is also the name of the American company that has developed the business intelligence tool. Presently, it is one of the best tools in the data science industry.

What is a Business Intelligence Tool?

Business intelligence tools provide methods and strategies to collate past, current, and future (predictive) data and generate useful business insights for data analysis and making essential business decisions. BI tools can easily handle a large amount of data from various data sources, and the data can be structured or unstructured. Using such tools saves a reasonable amount of time and resources, making them the need of the hour for today’s data-driven world.

Tableau does all of the above! Here are some highlights of the popular BI tool:

1. Drag-and-drop functionality.
2. Import data from multiple resources.
3. Present data in visual form.
4. Custom report generation.

Why Tableau? Is it Worth Learning?

Do you remember working on Microsoft Excel? Yes, you most probably have used it for creating an invoice or for, reading tabular data, or finding the sum of many values. Excel works perfectly for a small amount of structured data; however, for vast volumes of data being generated every day, a more robust tool like Tableau is superior to Excel (You will see how later in the article).

What is Tableau?

Although it is a BI tool, more specifically, Tableau is a data visualization tool that can generate charts, reports, and other visualizations. Basically, it can transform a lot of data into a comprehensible format without the need to have any technical skills or programming knowledge. It is an ideal tool for business analysts. Anything that is displayed as visual just registers in our minds better than text.

Further, Tableau can calculate all the fields like average, median, mean, standard deviation, variance, etc. For more complex charts, you may need some knowledge of writing simple queries. For example, sorting based on multiple conditions or grouping data on the basis of different factors. Still, for the most part, you will not write anything. Just drag and drop!

Apart from the fact that Tableau is flexible, secure, fast, and powerful, here are some features that give it the edge over other popular BI tools :

• Highly interactive solution with real-time data analysis and data blending.
• Requires no coding or technical skills.
• Easy integration with R and Python.
• Can combine data from multiple data sources. Also, able to handle large data sets.
• Easy to learn and fun to use.
• Mobile support is available.
• Interactive dashboard.
• Enables faster project completion.
• Each version comes with new capabilities.
• Great community support and superb documentation .

As per Gartner ratings, Tableau is as one of the leaders in BI and analytics for the last six years. See the figure below:

Still not sure about exploring Tableau? Well, what we have in this article will undoubtedly convince you to learn and use Tableau. Continue reading further.

Tableau Product Suite

Here is what you get in the Tableau Product Suite:

1. Desktop
2. Public
3. Online
4. Server
5. Reader
6. Prep

1. Tableau Desktop

Tableau Desktop allows you to view the story in your data through many options conveniently. You can drag and drop rows and columns to see different combinations of data and work on it to arrive at new trends and outliers visually.

Also, you have a plethora of options in the dashboard. This includes connecting to a data source from the list of available sources (that can be Excel sheets, MySQL, SQL Server, Oracle, data warehouses, cloud databases, etc.), mixing and matching, dragging and dropping from any of the sources into a single view.

You can perform various actions like creating maps, finding out the sum and median, or using filters too. Basically, Tableau lets you play with data in any way you want to. You can also switch between different charts (for example, a Treemap to a histogram and vice-versa) and find the most useful view for the current data set. The business intelligence software also finds trends and forecasts in the data.

All you have to do is click on the necessary options. Multiple people can access the interactive dashboards through the Tableau Server or Tableau Online. There are two versions of Tableau Desktop:

1. Tableau Desktop Personal – Offers limited access, and work can be shared only through Tableau Public. Also, workbooks are created in private mode.
2. Tableau Desktop Professional – Full access to all data sources is available, and work can be published on Tableau Server. It is ideal for users in the same organization who want to present their work at various levels.

2. Tableau Server

With Tableau Server, you can share your workbooks, and multiple users can access it through desktop or mobile from anywhere, any time. It offers a secure environment. This helps businesses create insights and answer in-depth questions on the go, making project execution faster and easier. You can load different sources and extract data from these sources or use live data for your analysis.

Tableau Server ensures you are using the right data for your review. It has options to inform others if a particular data set or field is obsolete for use through a simple message. By creating colorful visualizations, you can focus on the areas that need improvements for your business to grow further.

Whether you drag and drop fields to get your answer or type your query in natural language, you will instantly get your answer in a visual form. The server enables data management, like adding or removing data sources, permissions, views, and subscriptions, and thus enhances manageability, security, and scalability.

3. Tableau Online

Tableau Online is a SaaS that you can use anywhere without the need for hardware installation. You can explore, modify, visualize, and create insights through the browser, tablet, or even mobile phone. Also, you can access different data sources and connect to them on-premises or through the cloud. You can also add as many users as you want to access the workbooks that you create and work on the data together.

This is a compelling use of technology that enables you to perform reliable and quick analysis as a team. As already mentioned, you can also create your work on Tableau Desktop and then share and collaborate with others through Tableau Online to allow them to access your work and ask their questions about the data from different perspectives.

4. Tableau Reader

This is a read-only mode where a user with whom a workbook has been shared can view the visualizations and apply some basic filters without modifying the results. Workbooks can be shared through Tableau Desktop or Tableau Public.

5. Tableau Prep

This is a powerful tool that helps you view all your data in a single window and helps you clean and prepare data for visualization. With Tableau Prep, you can combine, reshape, reorganize and clean data from multiple sources visually, drag and drop the parameters that you need to prep(are) your data, apply aggregate, union, join, and other operations, and more.

You can also use calculations and see the results of joins and other query parameters to present data in different ways. Outliers and data inconsistencies can be fixed within minutes with this tool.

Moreover, you can view the changes as you perform them. In case you think you made a mistake, you can go back to any previous versions with a click. You can also keep track of the workflows, schedule them, and receive alerts if something fails. The flows are integrated so that you can monitor and share the data across your organization through Desktop or Online.

6. Tableau Public

This is the version we will use for our example. It is nothing but the public release of Tableau Desktop. However, the workbooks are not secure and cannot be stored locally in Tableau Public. Instead, they are stored in Tableau’s public cloud. That means even you can access the books shared by others and vice-versa. Most learners and first-time users use Tableau Public to get a hands-on experience with Tableau.

[An Example]How Does Tableau Work?

Okay, enough for the theory. Let’s put it to practice now:

• Download Tableau Public and follow the on-screen instructions to install it.
• After the successful installation, open Tableau Public.
• Next, connect to a data source. As we know, the data sources can be anything, an Excel sheet, CVS, SQL server, Hadoop file, etc.

Example Dataset

We have created the following dataset (with limited data) using Excel for the Tableau example. However, you can explore the existing sample datasets, too, if you wish to. Now, let's look into the data:

Out of the total rows, few are shown here. The data shows the names of students who joined on a particular date, their location/birthplace, average grades, average marks, and other information. This data will be seen in the 1st tab of the Tableau tool. Once you load the data, you can move it to sheet1. This is where all the magic happens.

In the above example, we have generated a stacked bar that tells us about the average marks by male (m) and female (f) students. The parameters in green are measures (numbers like average, sum, height, and weight), and those in blue are called dimensions (location, name, grade, date, etc.). For this chart, we have to pull the gender dimension to the Marks tab and average marks measure onto rows (change the calculation from SUM to AVG) as shown below:

We can drag and drop any size or measure depending on the results that we need and switch to any other type of chart as required. The bottom right shows the number of steps or dimensions, or measures we need for every kind of table to be generated. The above graph displays continuous value; however, we can also show discrete value:

Knowing the Grades of Students

Let us say we want to know the grades of each student based on their average marks. For that, all you need to do is select the average marks row, drag and drop category and name columns, select the chart of your choice, and bingo!

You can see that the data is neatly grouped based on the grade. The left diagram shows a histogram, while the right one represents Treemap. If you have data to practice with, you can keep adding more columns to increase the complexity. The more data you have, the more fun you can have playing with it.

For example, if we add a country column in our sheet, we can group students based on the country and get their median or average marks. Same way, we can also add a column for primary subjects and then analyze data on which topics are the most scoring for the students and so on.

Knowing the Students Scoring 80+ Average Marks

Let us say we want to know students who have scored average marks above 80 and in which subjects. For that, we need to create a filter:

Once we drag the screen, we are presented with something like this:

Important things to notice here:

• Data for all the students is shown; the ones who have average marks less than 80 are grouped under false. If you now go back to the data source tab, you will see this calculated field (avg_marks_above_80) added as a column with values true or false.
• The color coding indicates the location they are from. For example, pink is for Germany, orange is for India, and so on. Tableau puts default colors, but we can change them. Check out the birthplace dimension on the Marks's card.
• If I want to see only the records that have the false value, we can edit the filter and uncheck the true checkbox.

Country of Students Scoring 80+

Same way, if we want to know what country do the students who have scored above 80 in each subject belong to, we can drag and drop those columns:

You can drag the country in the Marks card and apply colors like this:

Based on the country or region name, Tableau automatically calculates longitude and latitude and displays the same in two different formats. We can represent the same using other charts like stacked bars or side-by-side bars. Let us get this information in a side-by-side bar by doing the following:

• Drag the average marks and significant subjects to the rows.
• We have already added a calculated field named avg_marks_above_80, and we are categorizing the information based on birthplace. Drag both as columns.
• Edit the avg_marks_above_80 filter to display only True values.

Now, you will get a chart like this:

You might be wondering why all the values look almost the same. This is because the variations are small in the data. We can start the index from 80 instead of 0, and then we can appreciate the differences more. To do so, double-click on the axis (average marks), and a window will appear:

Change the range from automatic to fixed and specify the range. Move to the next tab (Tick Marks):

Again, change the values of Tick origin and interval as desired. Now, we see the variations more clearly as follows:

Try creating another calculated field with average_above_90 and see if you get the following:

Note that these examples have been created for learning purposes only so that we can understand more features of Tableau. You can play around as much as you want by moving the columns and rows up and down and seeing how the charts change.

Let us Learn a Little About Line Charts

Tableau also shows you the parameters you need to generate a line chart:

Let’s say you want to calculate the average marks obtained by the class based on the year of joining and the primary subject. Sounds a lot. Tableau will do it all for you. All you have to do is drag and drop all the rows and columns you need.

Take the blue line (English), for example:

• As per our data, there are no students who took English as a significant in 1983 and 1986.
• In 1984, the average marks of those who had English as a significant were 84.9, as pointed out in the chart ((93.6 + 76.2)/2). If you hover your mouse on the point, you will get the details like this:

The same holds for other subjects too. All of this is just a representation of data. The real power of Tableau comes with analysis using trend lines and forecasts. The tool uses AI to show trends, discover patterns, and predict the future. To predict or forecast, we need data for a more extended period.

Insights from the Data

Let’s try to get some insights from this data. Click on Analysis, Trend lines and then Show trends:

Tableau will show something like this:

We can infer that while English and Science show a general upward trend, Math shows a downward trend. But this data, as we said, is not enough to generate the forecast. Thus, we have added some more data of students with different years of joining. Let us get the same line chart for the new dataset:

The data we have now is for 10 years instead of 3 years. Let us draw the trend line now:

As we see, with more data, we have better trends compared to previous trends. We see a downward trend for English, whereas Math is almost stable (no steep trend). If you click on to describe the trend, you can get all the mathematical values that are needed for in-depth analysis:

Let us now generate a forecast for the next 3 years:

If you don’t like the crisscrossing of lines, add major subjects as a row, and you will get individual lines like this:

Just for fun, switch from line chart to area chart to see the same results:

Now, we can see the description of the forecast:

Notice that for Maths, the prediction quality is still OK, but for others, it is not so. This could be because the data is insufficient or not of good quality. The forecast also shows a negative trend for English. The critical point here is how detailed the analysis is without us putting any significant effort. Every calculation was done by Tableau within minutes. Thus, we could swap rows and columns, add or remove columns, and play around with data to see different results.

Summary

This article is not exhaustive and does not cover every feature and option of Tableau. Tableau 10 Business Intelligence Cookbook and Tableau for Dummies are excellent books to start your journey with the popular data visualization tool. Along with that, you can watch some excellent tutorials on YouTube that teach loads of features of Tableau.

For the most part, though, you wouldn’t need a detailed tutorial. Moreover, you can practice Tableau to enhance your data science skills as well as to fulfill your enterprise’s needs.

People are also reading:

• Best Data Science Books
• Data Analyst Roles & Responsibilities
• Data Analyst Salary
• Top Data Science Tools
• Data Science Applications
• Data Scientist Salary
• Best R Interview Questions
• Tableau vs Excel
• What are Statistics and Probability?
• Data Analysis Techniques