Want to Build an Effective Dashboard? Think About Your Dream Home.

With hundreds of tools now available, data visualization has never been easier. Business intelligence tools like Tableau and Microsoft BI consistently release new features to make their interfaces more user-friendly. You do not need a background in data analytics to create effective dashboards.

After personally designing over 500 dashboards for internal performance tracking and external client reporting, I’ve found that building an effective dashboard is like building a house and consists of the following six main steps:

Step 1: Site Preparation – Data Connection

Before we can begin construction, we first need land to build on. Some factors to consider might be:

• Size of land
• Convenience of location
• Restrictions on the property

Similarly, you need to set up a data connection, the “land” on which your dashboard is built. Considerations might include:

• Size of dataset
• Data format and compatibility
• Restrictions on the dataset (null or missing values, manual refresh processes, etc.)

The land is the canvas upon which you start your project. For proper construction, the land must be leveled. Similarly, your data must be standardized before you can hope to extract any insight. It is important to not rush into using your first or most readily available data set. Carefully consider restrictions on your dataset to have the most efficient data setup. Chances are you don’t want to start construction only to realize your land is sinking!

Step 2: Foundation Construction – Data Cleanup

Now that you’re happy with the land, it’s time to start the foundation. Without a strong foundation, you might find yourself never finishing construction!

One common example is field name mismatches from combining multiple data sources. It is quite common that the same field will be named in more than one way (or one may be capitalized, spaced differently, etc.).

The solution? Use field mapping to help clean data quickly and efficiently. In my experience, good field mapping can make time spent on data prep and cleanup over five times faster.

Step 3: Framing – Metrics Selection

Now that you have clean data, you’re off to creating your beautiful viz!

This part has been greatly simplified for BI tool users in the past couple years. Understanding the difference between a dimension and a measure, difference between discrete and continuous data sounds simple, yet it isvery essential. A typical example is a Year to Date/Month to Date Year-Over-Year Growth chart. Knowing techniques for creating groups/sets of values is also useful to obtain efficacy.

Step 4: Installation of doors & windows – Ways to Visualize Data

The next step is to install your framing, doors, and windows. This step is vital as it will have a significant impact on the efficiency and the look of your house (i.e. amount of light, air circulation, etc.).

The way you choose to visualize data serves the same role as your doors/windows setup. You can always display data in at least three different ways, but which one will bring the most sunlight to your visualization/dashboard? Check out DOs and DONTs for common chart types.

Step 5: Roofing – Answer a Question

You’re almost done! Roofing is just like the Q&A process. A beautiful dashboard is useless without being able to answer questions! Imagine living in a house without a roof!

Step 6: Decoration – Make It Pretty!

Now that your nice, clean, and meaningful dashboard effectively conveys information, you can take some extra time to “beautify” it. Check out some formatting tips and tricks to make people fall in love with your dashboard!

Keep in mind: 

While designing your dashboard, consistently remind yourself, “What is this for?” This underrated question is often skipped, yet is critical to differentiate an effective dashboard from the typical unorganized monstrosity of charts and graphs currently dominating BI.