We learned in previous section that visualization is the actual output which end users will be seeing. Considering that visualization helps users get insights into their data it becomes important to build more intuitive visualizations considering the best practices.

Let us have a look at few of the good practices

• Understand the data: It is not possible to present data with meaningful insight if you don’t understand the data. You should be to make the business user get the best possible information from dashboard especially when there are multiple datasources. You should also know the nature of the data, like whether it is a historical data, a nominal data, or a ratio data. This will help you to understand the best visualization to choose and this enables business quick insights. Thus, it is always important to understand your data before transferring it to dashboards and reports.

• Use KPIs properly: We have seen that KPIs provides important information which gives users a quick insight into their data. If you use a single expression in KPI objects, they  show you the values. This KPI objects are not suited for situations where the need is to compare values with a benchmark and find out whether business is going in right direction or not.

  Thus, it is important that along primary value, you also provide secondary information. This helps users compare and decide if it calls for attention or not. You can also use the color code indication along with the KPI. For example, if you are showing a KPI like target achievement, then you can use color codes in the following way: if the value of the KPI is below range then show it in red color, if it is within the range then show it in amber color, or else it should be in green color. . This is because colors can easily get our attention.

• Choose color wisely: Colors are very powerful indicators; they can add value to the visualizations. Colors can easily draw our attention to important information and tie our thought process with other elements. For example, the bright colors can be used to get immediate attention.

  Colors also communicate the feelings. For example, red color means that something needs immediate attention, whereas the green color gives a positive feeling, that everything is going well.

  You should always limit the number of colors that can be used in your graphs. Instead of using 10 different colors, you can choose 5 best colors and use the different shades of those colors if required.

  While coloring graphs like table or pivot, make sure that you use colors only to show the outliers, rather than coloring all values of the chart. This will make it more readable and help in finding the anomalies in data.

  Use the light colors wherever possible. You may also ask your client about their corporate colors. This will make it easier for them to accept the dashboards more quickly as they will be comfortable with their own colors combinations.

• Less is more: Most of the developers make the mistake of developing many charts and graphs on a single screen. They feel like giving all the information to the users on single screen. But this approach may not work because too much information will make the users lose interest.

You should use minimum charts on a single screen and make your screen look cleaner. At max, you should draw four to five graphs on a single screen. This will help you to keep the users' focus on the most important chart and its information.

You can leverage the functionality of Alternative dimension available in couple of charts to change your dimensions from one to other. This will help you to reduce the number of charts required. For example, if you want to show the sales amount by various dimensions, like region, product group, sales person, and so on, you can put them as alternate dimension and change the chart dimension as required on the fly.

Let's take a look at an example of bad visualization and good visualization. Following image shows the bad visualization:

You can see in the preceding dashboard that though the KPIs which are selected are best for the dashboard, the presentation is not good. Proper chart types have not been used to show correct information, and also the size of the chart is wrong; for example, the pie chart is totally impossible to be understood.

Now the same KPIs can be converted into a good dashboard, as shown in the following image:

As you can see, it is much clearer than the earlier dashboard. It presents correct information to the end user, starting from showing the overall KPI, then its breakup, and then the trends to give additional information.