Here are 10 easy tips to let you get the most out of your Tableau Software visualization effort.
- Use Excel to get your feet wet. Rather than start working with a database, it's usually easier to get started with some simple Excel spreadsheets. You'll learn how to use Tableau much faster this way, because all you have to do is make a quick change to the Excel data in order to see the effect it has on your Tableau visualization. Keep that "make a change, save, view the change" loop tight!
- Clean your data. Your data might have mixed upper-case and lower-case; it might have missing values; it might have mixed formats. (This happens with Dates a lot!) For Tableau to do its job, the data needs to be as clean as you can make it.
Usually, you can do most of what you need to do directly in Excel, by using its data formatting tools. Sometimes, you might be forced to play tricks by exporting your data to a CSV format, and then clean it up either by hand, or by using automated tools (see for instance Google Refine). If you have some scripting talent, you can also write some pretty simple Python scripts to clean up a data source. I've found that it rarely takes more than a couple of hours to clean up data, once you identify where the problems are. - Normalize your data. The big issue here is that often you'll start with a spreadsheet that has more than one column representing Date or Time. For example, you might have a spreadsheet simulating a calendar, where it has a column representing each month in a year. You're going to want to have no more than two columns that represent Date or Time in any one worksheet! (I say "two" because it is OK to have a "Start Time" and "End Time", in data that models events with duration. Even then, you might be better off with "Time" and "Duration".)
Secondarily, your spreadsheet might have rows of repeated data--so-called "grouping". In this case, the right thing to do is to create two spreadsheet tables in separate worksheets. If data is grouped or repeated in a spreadsheet worksheet, you need to split that worksheet into two or more!
Make sure that if you split your data, that you leave a "bread crumb" from the detailed worksheet to the "group names" worksheet. Usually, the bread crumb can be as simple as adding a column in the detailed worksheet that identifies the name of the group. Tableau will then be able to link the two worksheets and "do the right thing." - Be aware of demographics data on the Web. There are lots of nice public sources of data on the Web. The great thing about them is that they let you use your data in ways you might not have imagined.
The canonical example of this is calculating some kind of result per capita. Per capita census data is readily available on the Web. You can start from, say, here.
Once you load census data into your Tableau workbook, you can use either data joining or data blending to create far more meaningful results. - Be aware of various sources of maps on the Web. There are many interesting maps on the Web. You don't have to satisfy yourself with the one that Tableau provides for you. It is very easy to load a map into Tableau, and tell it where its "origin" latitude and longitude are on the map. This can make your visualizations look really impressive, depending on your audience.
- Be aware of useful analytics tools for Tableau. You can use these to do things like regression analysis, without doing any deep coding work. The Tableau community has made many of these functions freely available.
I especially like Joe Mako's analytic contributions to the community. You can find a list of his contributions here. - Create dashboards that support a very high-level view (with a map), and at least one lower-level view. This "drilldown" paradigm is great, because it creates confidence in your customers' minds. If they can see how the aggregated data (at the top of the screen) correlates to the detailed data (at the bottom), then they think there's a decent chance they can trust the data. And once they start trusting the data, they might actually start using the data to make decisions they wouldn't have considered before.
- Learn how to embed your Tableau visualization into Microsoft SharePoint. The trick is using SharePoint's Page Viewer Web Part. Add the web part into your SharePoint page, and use the PermaLink that Tableau creates for you. It's very, very easy. Best of all, it's an argument you can easily make to people who ask, "Does your shiny new tool work in SharePoint?"
- Use Tableau Public when you need to create a demo, right now. Once you have a bit of experience under your belt, you'll find that you can create an extremely impressive demonstration with no more than a few hours of work--sometimes far less, if your data is in decent shape to begin with. The key is using Tableau Public, which puts your visualization directly on the Web. (Of course, you'll need to be careful not to put anything too sensitive out there.)
- Use Tableau's knowledge base. Tableau has put together many great "how-to" videos. You will save yourself a pile of time by reviewing at least 10 of them, before attempting your first test.