[Part I] Tableau Features— The Complete Learning Guide
Tableau is a leading data visualization tool used for data analysis and business intelligence.
In this article, I walk through all the details and important features of Tableau to get you from zero to hero. You can download Tableau here.
As Tableau is a visualization tool, you must connect to a data source(s) or table(s) to get started.
Relationships
Think of a relationship as a contract between two tables. When you build a viz with fields from these tables, Tableau brings in data from these tables using that contract to build a query with the appropriate joins.
A relationship can be established between tables when working with Tableau, and the cardinality configuration (e.g., many to one) should be done properly. The default is Many to Many relationships but it can lead to bad performance (when it is unnecessary). And when the cardinality is configured to One Tableau stops the scan once it finds a match. So improper configuration can lead to miss-performance or wrong results.
Using Relationships at the logical layer rather than a join in the physical layer is recommended.
Data Blending
Data Blending uses the logic of a left join and combines data from two data sources at the visualization layer.
Data types — Numbers & Strings issue
In some cases, Tableau can assign the type string to decimal columns (e.g. sales). Changing the type directly from string to number will result in having a null in the values of that column. So instead what you need to do is to change the Locale to how decimals are represented (with , in non non EU countries or with . in EU countries)
Measures vs Dimensions
Measures enable data aggregation (e.g. SUM, AVG) and are usually numbers while dimensions are usually strings or the numbers in case of IDs.
To create a report follow the rule Measure by Dimension (e.g. sales by product).
Creating a Dashboard
After creating the sheets (e.g. Sales by Country) you need to drag and drop the different sheets in the dashboard.
Filters: Continuous vs Discrete
With a discrete filter (for the categories of products) we can select and deselect values and with a continuous filter (for the SUM of sales) we can have a range with a slider as shown on the right.
Detailed analysis with discrete values
Trend analysis with continious values
Hierarchy
Hierarchies can only be created with dimensions and not with measures. They help organize views and explore data at different levels. When doing the visualization, you can drill down or up in the hierarchy by clicking on the + or — icon in the Rows field.
Another example of the hierarchy is the one for the Date dimension which is created by default for dates in Tableau (this is not the case in PowerBI).
Year > Quarter > Month > Day
Cluster Groups
Tableau adopts the K-means clustering as the clustering method. This will help to better classify the data and interpret it.
Clustering the top-selling products by clustering products based on quantity and profit. To do this graph, we have two measures (profit & quantity) and we drag and drop the Dimension product ID or Product Name on the detail Mark. Then to do the clustering, move to the Analytics pane.
Histograms & Bins
We defined earlier that most of the time we create visualizations measured by dimension (e.g. Sales by Product), now we want to make a graph of a measure by a measure which consists of turning a measure into Bins.
One important thing to note is to convert the (score) bin which is a dimension to a continuous measure in order to have a histogram.
To interpret this histogram which is skewed to the right to see the distribution of the customers based on their score. We can say that many new customers have low scores.
I hope this guide helped get you started with Tableau and learn practical knowledge on how to use it, don’t miss the next part ;)
