What Are Sets in Tableau and When Should You Use Them?

This article is to provide a quick overview of what sets are and when they are most commonly used. 

Firstly, a set is a custom field that you create in Tableau Desktop that represents a group of members in a dimension matching a specific criteria. For example, if you have a map of sales revenue by state, you could very quickly shade the map for states that are above x-dollars in sales using a set, where x represents a user-defined threshold of sales revenue. For readers that are familiar with statistical concepts, this is the same as cohort analysis.

The three main uses of a set are:

• Create a subset of the data – select one or more dimension members that are of interest you. This sales threshold example above represents this usage. However, sets can be static as well, so that the members of the set are manually selected by the user and will not change over time.
• Create unique encodings – combine dimension members to create unique encodings. This approach allows the user to create a set that combines two dimensions into one. E.g. Product and Region. A set of Product and Region would create a cross join of the members from both dimensions, so that a user could quickly filter out combinations such as West, Furniture; or East, Office Supplies; etc. 
• Save filters for a later use – once you have created a set, you can treat it like a filter to only show members of that set within your view. However, unlike a quick filter, sets don’t have to be recreated for each worksheet because they are saved and are accessible across all worksheets within the workbook.

Although sets cannot be referenced directly by secondary data sources, it is possible to reference such sets with the use of calculated fields. This method can be very helpful when wanting to filter down a view from one data source with a set existing in another data source. This mimics a “blended set”, even though sets cannot be used directly by different data sources from which are originally based off of.

For a video tutorial on sets, check out the following resource:

http://www.tableau.com/learn/tutorials/on-demand/sets

How to Pass the Tableau Desktop 8 Qualified Associate Exam

Recently I took the Tableau Desktop 8 Qualified Associate Exam and passed it. Having taken other certification exams before, I knew that study guides issued by exam providers don’t always provide the best methods of preparation. As usual, I was unsure about the best approach to study for this exam, and ended up exhausting the resources I had available to me. All of the materials I studied were free and were sufficient resources to help me pass the exam. Below is a compilation of study materials on how you can go from not knowing what Tableau Desktop is to becoming a Tableau Desktop Qualified Associate with no training costs involved.

Learning Resources

• Official Exam Preparation Guide
• 14 Sample Questions
• Percentage breakdown of all skills measured as well as links to articles and tutorials on each topic
• Guided Video Tutorials
• On-Demand Training (~ 30 videos)
• Live Online Training (~ 8-10 1 hr sessions)
• 5-8 personal exercises. Get creative with the types of visualizations you find interesting!
• The Basic Tutorials on my site are good starting points to familiarize yourself with some of the following features in Tableau:
• Inner Joins, Left Joins
• Box & Whisker Plots
• Quick Filters
• Marks card (color shelf, size shelf, detail shelf)
• Basic formatting
• Excluding data points
• Dual Axes, Synchronized Axes
• Quick Table Calculations (Percent Difference)
• Continuous vs Discrete Dimensions
• Trendlines
• Annotations
• Editing Data Sources
• Editing Axes
• Building Dashboards
• Dashboard Actions (Filters)
• Move on to the Advanced Tutorials portion once you've mastered the above features!

Best of luck studying! Please reach out if you have any questions!

- TableauWiz

Why You Should Blend Instead of Join

Even to an experienced SQL query writer, the mechanics of blending data, rather than joining it, can be difficult to understand upon initial exposure. Data blending is a fairly advanced concept; this article explains the concept of data blending at a basic level using Excel as an example.

Data blending in Tableau is the operation of combining multiple data sources into the same view by finding common fields between them to join on.

How is this different from a regular join?

Let’s first describe the equivalent of a data blend using an example with Excel spreadsheets.

Suppose you have two different Excel spreadsheets, the first sheet contains data on sales, the second sheet contains quota for sales. If you wanted to do a side by side comparison of sales vs. sales quota by country, you would likely first aggregate the sales data to be summed by country, then perform a vlookup to the sales quota sheet on each country to grab the quota for that country and place the result next to the sales for that country in the aggregated table.

This concept is referred to as a post-aggregate join, as the vlookup to the sales quota spreadsheet did not occur until after the sales data was already aggregated by country. Therefore, the “join” to the second data source was done post-aggregation. This process of joining data sources post-aggregation is referred to as data blending in Tableau.

Had we performed a vlookup on the sales data before aggregating by country, we would likely have multiple sales records for each country with the quota next to each record. In order to get the desired comparison, we would have to aggregate the data by summing the sales column by country, and grabbing the max or min sales quota by country. This process is understandably more work, as two different aggregations have to occur, and much more data ends up being processed at once because the vlookup occurred before the aggregation.

Creating this in Tableau would mean having your sales data in one data source, sales quota data in another data source, creating an aggregated view of the sum of sales by country, and simply dragging quota onto the visualization next to the sales column. Tableau will automatically recognize the common field of Country between the two data sources, and use that to do a post-aggregate join. Tableau does not join the two data sources until after the data is aggregated. Given that the desired visualization aggregates the primary data source, the post-aggregate join is much better for performance than joining at the row-level first and then performing the aggregate calculations.

5 Reasons to Use Tableau Reports for Salesforce Data

Building reports and dashboards in Salesforce can often prove to be difficult due to the limitations imposed through the out-of-the-box functionality in Salesforce reporting. I have listed five existing limitations in Salesforce reporting as well how pulling your Salesforce data into Tableau to build reports and dashboards can address these limitations.

1. Calculated Fields 

Oftentimes, users will see a report and will want to create a new metric for it. For example, you could be looking at a report of your revenues by account and wonder what the percentage breakdown of total revenue is for each account. Getting this information in Salesforce would require creating a new Formula Field. In Tableau however, in addition to the calculated field option, you could simply right-click the existing revenue field, and change it to be a percentage of the total in just a few seconds. This is referred to in Tableau as a Quick Table Calculation. Tableau offers maximum flexibility with these calculations in terms of the desired level of aggregation. E.g. you can choose to get the percentage breakdown by row, column, or both. The quick table calculations are not limited to percentages either, as they include creating Running Totals, Moving Averages, Percent Differences, Ranks, and others. The available flexibility and ease of use in creating such measures is what makes Tableau the winner when it comes to creating calculated fields.

2. Connecting to Multiple Data Sources

A large limitation of Salesforce is that it can only quickly report on data that exists within Salesforce. Salesforce doesn’t have built-in connectors to point to external data sources such as an Excel file or a SQL Server database. To integrate such data, using more complex methods of integration such as web-service API’s must be employed. The data must also be stored within Force.com to build the report as well. Fortunately, the process of integrating data from multiple sources within Tableau is much simpler. Users have several dozen built-in connectors to external data sources in Tableau such as Amazon Redshift, Hortonworks Hadoop Hive, IBM Netezza, Microsoft SQL Server, and even Salesforce.com. The connections to these data sources can be live connections or extracts, allowing the user to decide whether to pull the data into Tableau for processing or leave it pointed to the external data source. This flexibility makes Tableau the BI tool with the most built-in external data source connectors.

3. Default Chart Types

Visualizing data with charts and other graphics can be essential to understanding it. Salesforce’s built-in capabilities provide a very limited number of chart types to visualize data. The charts available in Salesforce are limited to the following types: bar charts, column charts, line charts, pie charts, donut charts, funnel charts, and scatter charts. This means the standard user is not able to create a map to see how their sales are distributed across different geographies, or a simple box plot to see what the distribution of opportunity amounts are across different accounts. The out-of-the-box functionality to create bullet graphs, histograms, or bubble charts is simply not available with Salesforce reports. These limitations ultimately hinder the ability of decision makers to gain insight from the reports.

4. Interacting With Data

One of the most useful features a report can have is allowing users to interact with the data they see. This is something Tableau delivers on very well, and something that Salesforce requires a lot of customization for in order to implement. When looking at a bar chart of sales by region for example, a user may want to click into a specific region and see a new bar chart that shows the sales of the various accounts composing that region. This drill-down and filtering capabilities don’t come standard in Salesforce reports and dashboards, but are very easy to implement in Tableau. To implement something like this in Salesforce, Force.com APIs would need to be employed, preventing the average administrator from building interactive charts for users.

5. Cross-Object Reporting

This is a big one. Salesforce has difficulty reporting across multiple objects that are not parent/child objects but still have common data, making them worth combining into a single report. For example, to get insight into which lead sources are converting to opportunities and revenue, two different reports would need to be created and then merged in an external reporting system to gain insightful information from the data. Tableau is able to combine data from multiple sources onto a single report simply by recognizing common field names and data values. This feature in Tableau makes it possible to not only quickly create a report like the one just mentioned, but also to create reports with data outside of Salesforce. For example, if you had a monthly sales quota spreadsheet in Excel broken out by salesperson, you could easily overlay this data in Tableau with existing opportunity data pulled in from Salesforce to get quick insights into which of your salespeople are meeting their quota for this month, year, etc.

This article is not intended to bash Salesforce as a CRM tool, for which it is best in class. It is simply to state the reporting limitations of Salesforce and point out reporting-specific tools like Tableau which can address these limitations. Please share your thoughts on these points as well as any other limitations you have come across.