Visualizing Data: Part One
Video: Tableau Desktop
Tutorial Explanations
Common Visualization Task: Formatting Data
- Formatting your data correctly is critical for the successful completion of all other steps. Your necessary formatting depends on the tool you use. By and large all InfoVis tools will require some sort of formatting unless the data specifies that it's preformatted to certain tools (which I can't imagine will happen often; data is typically formatted to be read in chart-form easily).
- For broad-level formatting (e.g. you've got variable fields in rows instead of columns) there are macro tools available to format. This is especially useful large-scale datasets (the kind you'd likely be interest in visualizing), but it takes a bit of searching to find them. For Tableau (and, by extension, the format-similar Motion Chart) the Tableau Dev Team has created a Tableau Data Reshaper add-in for Microsoft Excel. Installation documentation is standard for Excel Add-ins and is included in a Readme file and is available online.
- For a tools like Tableau and Motion Charts, in which you might want to visualize differences over time, one record (row) for any given unit of analysis needs to be a devoted to the variable associated with a single year. In practice, this means you'll need to format your data so that your columns are: unit identifier (e.g. Name, Country, etc), a single year (or more specific dates may be acceptable depending on the tool), and then all the variable values for that unit for that year.
- Beware incomplete fields! Variables that lack data for all units of analysis will cause trouble with visualizations. The Excel Reshaper Tool, listed above, deletes those variables to obtain a completed dataset. This may influence the data you can use for your visualizations and may therefore restrict your ability to map the data you collected.
Tableau: Importing Data
- After creating your formatted spreadsheet, follow these steps to import your data into Tableau Desktop:
- Choose "Connect to Data" in the top left area of the interface, beneath the toolbars.
- Choose "Microsoft Excel Spreadsheet" from the options list.
- Browse to and select your file, then OK.
- In the Workbook Connection dialog that appears, choose the single table from the list if you only need one.
If you need multiple workbook tables, you must link them. This is explained visually in the video, but you simply need to select 'Multiple Tables' and then 'Add New Table...'; choose the second table and the fields you're interested in, then click the 'Join' tab and configure a Join Clause (choose a field from the first table that is identical to a field in the second table). - 'Import All Data' isn't required, but it allows Tableau to run more efficiently. If you choose to allow it, save the database and let it load the data into the program.
- If your data includes any standards in its Dimensions that Tableau supports, you'll want to verify it recognizes all entries. You can do this by right-clicking on the Dimension and moving to the bottom area of the context menu. There you can find options for Data Type and Geographic Role. Options are included in an extended radio list and you can choose 'Edit Locations' at the bottom to see all your values and which ones Tableau doesn't recognize. They can be corrected from here as well.
Tableau: Planning Your Visualization
- By this point you likely already have an idea what kind of visualization you're interested in creating. If it requires information such as a map, lay the appropriate measures (Longitude and Latitude for maps) by double clicking them. Tableau can auto-place common measures quite intuitively. Don't rely on this method for many other measures, however. If it isn't a common visualization base you can't count on it to be placed appropriately. You can click and drag measures to place them in the field or axis of your choice.
- If you aren't certain of a visualization, experiment by choosing the dimensions and measures that interest you and then choose 'Show Me!' from the toolbar. This will display a simple graphic tray depicting the visualizations Tableau can create using those measures (based on their perceived data type). Choose the option you prefer to load a basic-format version of that visualization type using Tableau's estimation of how to sort your variables.
Tableau: Creating Your Visualization
- Whether you've created a visualization with 'Show Me!' or are filling a base you've created, you'll want to customize which of and how your data displays. Drag the appropriate Dimension onto the main area of your visualization to set it as the focus. You'll know you've done this correctly if a 'Show me!' icon hovers near your mouse when dragging over the visualization area. You can add additional dimensions to the visualization by dragging more.
- Adjust the unit symbols by changing the Marks type from the dropdown menu in the Marks area of the visualization tools. This can be cosmetic (changing the shape of an identifier point) or add additional data dimensions (moving from a one dimension symbol to a multidimensional mini-chart).
- In the case of multivariate Marks, you can set relationships to display within your Mark. To compare measures, use the 'Measure Values' Measure to fill the field representing the relationship (e.g. 'Angle' for Pie Charts).
If you have several Measures you'll need to excise those that don't interest you. Do so by dragging Measure Values out of the 'Measure Value' window and back to the 'Measures' area. - Make sure to also set the differentiation (e.g. Color) to the 'Measure Names' Dimension. You can adjust the colors given to each Measure in the 'Measure Names' window. Put your mouse over the top-right corner of the 'Measure Names' window and a small settings icon will appear, containing an option to 'Edit Colors...'. You can choose from Tableau Palettes, but make sure you choose colors that are different enough to be easily seen as well as of a significantly higher saturation than the visualization background. Dull colors won't stand out preattentively. Also, avoid using low-saturation Marks on a high-saturation background. The eye doesn't pick out low-saturation areas pre-attentively and you'll lose much of the value of your visualization.
Tableau: Make Your Visualization Manipulatable
- Give users the ability to manipulate the data to add a far greater value to the visualization. Instead of mapping date as an axis measure, free up the visualization to measure other relationships and use date as a Dimension with the Pages function. Change a date measure like 'Year' to a Dimension by dragging the variable from Measures to Dimensions. Once there, simply drop it in the Pages window to sort the visualization data by year. One year's data will be displayed at a time, and users can "play" the data changes as well as set the speed of the animation by using the 'Year' control window that appears.
- Pages greatly enhance the value of a Tableau visualization by bringing animation into the mix. This adds pre-attentive fluency to variable shifts so that we associate that the moving Marks represent the same units of analysis or measures across time.
Back to Final Project | Tutorial Pages: Part Two