Tableau Server and Tableau Desktop
Tableau is a business user-friendly analysis and data visualization tool. Its core comprises a full client and a web server with connectors to a wide variety of data sources, including local data. The structured intuitive user interface, built-in intelligence and memory utilization – geared to optimizing performance – contribute to the popularity of this solution in self-service BI (SSBI) and data discovery scenarios.
Tableau Software (Tableau) emerged from scientific research at Stanford University, where its three founders pursued a vision of providing business users with software that allowed intuitive analysis and insights into data. The company has enjoyed remarkable growth worldwide since its inception in 2003. Since 2013, Tableau has been listed on the New York Stock Exchange.
For some years now, the vendor has invested increasingly in internationalization. Its corporate headquarters is located in Seattle in the United States and the EMEA headquarters is in London. Tableau Software has offices worldwide, with some of its branches still operating as pure sales offices.
Tableau offers four core products: Tableau Desktop, Tableau Server, Tableau Online and Tableau Public, which are all based on one common technology. This review focuses on Tableau Server and Tableau Desktop as a combination in enterprise scenarios.
Tableau follows a strategy of delivering software that requires as little training as possible, and allows business users to analyze data by means of visualization without having to rely on the assistance of analysts or even IT staff. At the same time, the tool offers predefined statistical functions and additional analysis capabilities specifically geared to analysts for in-depth analysis.
Like other self-service BI providers, the company pursues a “land and expand” strategy with a focus on business users. In addition to a free trial version of the Tableau Desktop client, Tableau offers a free software known as Tableau Public. Anyone can use Tableau Public to visualize data and publish it to the Tableau Public cloud. Visualizations can be embedded into other websites (e.g. for data journalism or blogging) and retain their interactive capability.
Tableau’s original products are Tableau Desktop and Tableau Server. Tableau Desktop is the main authoring software supporting the user through all steps of the analytics cycle (from data to insight), while Tableau Server is the web and mobile-enabled dashboard publishing environment. With this architecture customers typically have a select few developers and many more users (sometimes hundreds) who consume reports, dashboards and analyses. Furthermore, Tableau has a number of reference customers for deployments supporting thousands of users.
Tableau Online is the multi-tenant cloud solution, which is the hosted version of Tableau Server and therefore uses Tableau Desktop for the creation and deployment of analyses and dashboards. The product works with on-premises and cloud data sources, and is typically used for analysis scenarios without IT involvement or sharing analysis results with users outside the firewall.
Tableau has various different types of technology partners. It partners with database and application vendors such as SAP, Cloudera, Salesforce.com and others by building native connectors to these systems. Partnerships with companies like Informatica and data preparation vendors like Alteryx support customers’ data integration needs. To extend the product’s capabilities, Tableau partners with specialists such as Datasift to provide social sentiment analysis. Finally, for advanced analytics (e.g. cluster analysis), Tableau can connect to R, Python, SPSS and SAS data sources.
Tableau has a proven client-server architecture based on optional, proprietary columnar and optimized data storage. The core authoring component is the desktop client. Data preparation and modeling are done in Tableau Desktop. The software’s characteristic features include native connectors to over 55 data sources. In addition to relational and multidimensional databases, on-premises and cloud business applications can be accessed. Tableau is also capable of integrating local files. The capability to parse and import data from tables in PDF files was added in version 10.3.
Functions for parsing and automatically cleaning Excel files to speed up importing are included as Data Interpreter. There are functional differences when connecting to data sources due to the underlying language differences between MDX and SQL. As a result, some features are not available when connecting to OLAP data sources, but are available when connecting to relational data sources, and vice versa. Data source definitions can be re-used. They can be stored in a Tableau Data Source (TDS) file and shared with other users (Shared Data Sources) via e-mail, or published on the Tableau Server.
Tableau has two different ways to connect to data sources: “Live-Connect” leaves data in the underlying data storage and queries the data at runtime. In this case, Tableau acts as a pure front-end analysis tool and does not store the data. The second option is to use Tableau’s own data storage facility (Tableau Data Engine) to create a Tableau Data Extract (TDE) and to load data into memory. Using a TDE is beneficial when the underlying database is slow, if a user wants to reduce the load on the underlying system, or when a user wants to work with data offline. Tableau introduced a powerful hybrid connection in version 10 called Cross-Database Join: this feature allows data to be joined from different sources (e.g. local data crossed with enterprise/governed data sources such as Microsoft SQL Server or Oracle databases). To save time during data preparation, a smart join feature actively recommends relevant joins to users.
The solution saves analyses and dashboards in Tableau Workbook (TWB) files. Another file format, TWBX, stores data and visualizations together in a single file for offline consumption. Offline read access is possible via a free Tableau Reader. Tableau Reader is similar to Adobe’s PDF Reader and presents users with analyses and dashboards in an interactive form.
Tableau Server is the solution’s application server and acts as a central repository and portal for all Tableau metadata concerning users, data connections and content. Tableau Server administrators can perform different tasks such as scheduling data refreshes, managing users or published content and so on. The metadata is stored in an internal PostgreSQL database. Tableau Server also allows users to consume and interact with and even modify published visualizations on their web or mobile clients. In addition to Tableau workbooks, dashboards and views, users can also access published Shared Data Sources to create new analyses. In version 10.3, Tableau introduced features for data-driven alerts and custom content subscription.
Most Tableau Desktop users are non-technical business users rather than data scientists or SQL experts. Data access and modeling are performed in Tableau Desktop using a graphical interface with built-in intelligence to assist users. For instance, any two identically named fields are automatically connected or recognized on the basis of their data type or format, and declared as measures or dimensions. Tableau automatically recognizes dates and creates a hierarchy from year to quarter to month to week, and so on.
Tableau’s primary focus and forte is interactive and intuitive visual analysis to efficiently dissect data and find patterns. Entities such as measures and dimensions are selected via drag-and-drop for analysis and visualized instantly in interactive visualizations. Users are supported by built-in intelligence. For instance, a “show me” function provides proposals for appropriate visualizations for any combination of data fields selected by the user.
Besides the creation of hierarchies, the tool offers additional grouping and calculation options supported with context sensitive wizards. For example, a user can select entities on a visualization to create sets and therefore group common features or attributes for further analysis. Tableau’s calculation library’s analytical feature set ranges from simple trend and linear analysis to non-linear analysis, forecasts, statistical quartiles, standard deviations, table calculations, and other analytical and statistical functions. Cluster analysis was added as a predefined statistical function in version 10.
In Tableau, visualizations reside in individual sheets much like the sheets in an Excel workbook. To create a dashboard, the user drags the desired worksheets from a shelf onto the dashboard canvas and re-arranges the visualizations. In addition, users can extend the information on a dashboard by linking to further information on the web via a “URL action”. Tableau also possesses a parameter and global filter concept. Parameters allow dashboard consumers to filter the data displayed to a specific range.
Visualizations and dashboards allow focused display on the screen. Tableau does offer very limited functions to specify page formats and control printer output. Printing takes place according to the WYSIWYG principle. Workbooks can be printed out or exported to PDF, Excel and CSV, distributed as files via email or URL, or embedded into a website or portal.
In Tableau Server, users can collaborate and discuss content via comments, annotations and dashboard web page objects. Fields can have comments that describe them. Annotations can be used to call out a specific mark, a specific point such as a value on an axis or a reference line, or an area such as a cluster of scatter marks, which is important especially when compiling data stories in the relevant storytelling section. In addition, conversations and annotations can be shared on Tableau Server alongside the applicable visualizations and dashboards.
User & Use Cases
The use of Tableau is shifting from small to larger enterprises and from smaller to larger deployments as well.
Correspondingly, the use of basic data analysis is declining and the consumption of pre-built content remains high (dashboards) or is increasing (standard reporting). But still, Tableau is used to prepare and analyze data more frequently than the average of all products in The BI Survey.
Current vs. planned use
N = 75
5 products most often evaluated in competition with Tableau
N = 72
Percentage of employees using Tableau
N = 75
Number of users using Tableau
N = 75
Tasks carried out with Tableau by business users
N = 70
Company size (number of employees)
N = 75
Want to see the whole picture?
BARC’s Vendor Performance Summary contains an overview of The BI Survey results based on feedback from Tableau Server and Tableau Desktop users, accompanied by expert analyst commentary.Purchase the Vendor Performance Summary