Getting Started 4: Building Visuals

Published on March 14, 2018

Arcadia Instant and Arcadia Enterprise are our company’s tools for visual analytics. Arcadia Instant is a free downloadable visual analytics tool while Arcadia Enterprise is an enterprise-grade platform capable of supporting teams of users and variety of data sources.

Arcadia Instant comes in Windows and Mac downloads directly from Arcadia Data, and helps you quickly gain insight into your data. So, you can explore and visualize your datasets with interactive visuals, dashboards, and apps.

Arcadia Enterprise can be installed yourself with Cloudera Manager or Ambari Stacks. You can find detailed directions on our documentation portal.

For this quick tutorial, we will go through the detailed steps of building visuals and settings that affect how a visual is display and behave.

 

Overview

It could get a bit confusing on the different terminologies used so let’s do a quick review of how visual, dashboard, and apps relate to each other. As we already discussed, visuals are individual graphs composed of the information in the dataset, or from external sources. Dashboards are collections of several related visuals on the same canvas, and may or may not have built-in interactions, such as click behavior and parameter passing. Apps are named collections of several dashboards and visuals, which are often related. We can define apps to contain multiple tiers of dashboards and visuals. When we externalize apps, we call them standalone apps.


Creating Visuals

Now let’s start building a new visual! There are many ways of creating a visual:

  1. Create a new dashboard first – When you are starting out, you can create a visual by first creating a new dashboard then create a new visual by selecting the connection and dataset inside the dashboard.
  2. Create from dataset – You can also create a visual by going to Data and selecting the New Visual icon.

Let’s use the first method to start creating the visual.

  1. Click on the New Dashboard to be our new visual’s container.
  2. Now that we have a Dashboard to put the new visual in, let’s start the visual by selecting the new visual by selecting the connection and dataset from the right side. In this case, we will select the default, which are sample and Population Census.
  3. Now we see a table visual is created in the Dashboard, To edit it, we need to hover over it and select the edit icon on the corner of the visual.
  4. At last, we are now in the visual designer! There are many features in this interface, which varies by the type of visual selected. The default visual type is table. Some of the features are common to all visuals, so it is worth the time to take a closer look.
    1. A visual has an editable title at the top of the page
    2. A visual also has an editable subtitle that displays under the title at the top of the page to provide additional information to the user
    3. The visual canvas shows the current state of the visual.
    4. The visual type menu allows you to select different visuals, such as bar, line, maps, and so forth.
    5. The shelves, in this case, Dimensions, Measures, Filters, and Limit, are contextual, which means they are dependent on the selected visual. We define the visual by dragging the field dimensions and measures onto the appropriate shelves.
    6. The available fields are defined, including the Dimension and Measures of the dataset. These are also contextual.
    7. The submenu lets you switch between the other options available to configure this visual, such as defining the look and feel style.
  5. For our first visual, let’s build a bar visual from our dataset. Click on the bar option in the Visuals menu.
  6. Notice that the shelves changed: we now have a mandatory X and Y, as well as the optional shelves, Colors, Tooltips, and Filters.
    X and Y are for the principal dimension and measurement, respectively. Drag the dimension State onto the X shelf, and the measurement Population onto the Y shelf. Notice that on the shelf, Population appears as sum(Population). This is because, by default, the Y shelf shows aggregations of measurements. Also, our dataset defines default aggregation for each field; for Population, it is the sum() function. To see this basic bar visual, click Refresh Visual, just under the shelves.

  7. Your first visual is somewhat informative as you can see which states have the lowest and the highest populations. However, there is so much more we can do! However, let us name and save the visual first.
    • Click Edit next to Untitled, and change the title to Comparison of State Populations.
    • Click the big green Save at the top left of the screen.
  8. Notice that the visual we have so far is ordered alphabetically by state. That is nice, but if I order it by the size of measurement, it becomes easier to pinpoint the categories with the largest and smallest values. To do that, click the side arrow next to sum(Population) on the Y shelf, and selecting Order and Top K > Descending from the nested drop-down menus.
  9. After clicking Refresh Visual, we can see the sorted bar chart. This display is much more informative.
  10. However, it is still difficult to see individual states. Arcadia has several solutions for that, and one of the most useful ones is the Color shelf mentioned earlier. So, let’s drag the State field onto the Color shelf, and click Refresh Visual again.
  11. Hover the mouse over any of the bars in the chart; you can see a tooltip with some basic information. This comes directly from the fields on the shelves. For example, it will show that Texas (abbreviated TX) has a sum(Population) of 130M. This is something we can extrapolate from looking at the graph, even if the exact figure is not available. Still, if you add more information to the Tooltip shelf, it provides the user with in-depth detail.
  12. Add Record Count to the Tooltips shelf. Incidentally, Record Count is not part of the base data table; you will not find it if you look at the raw data. Instead, it is one of the “ease of use” improvements that Arcadia generates as part of the dataset definition. It counts the number of records that satisfy the query.
    After clicking Refresh Visual, hover over one of the bars again to see what information it provides.
  13. It looks like there are 14 individual records in the sum(Population) aggregate of 130 million people. This makes sense because we know that while Texas is large, it is not quite enormous. The value of 130M represents the sum of populations obtained from 14 separate US Census measurements, starting in the year 1790 and ending with 2010. All well and good, but now we see that the graph we built communicates nonsensical information. However, it would be very useful to see the population of each state for a particular Census. If only we could filter our data on a particular value.
  14. Arcadia to the rescue! Drag the Year field to the Filters shelf and you will be presented with a list of years to select from. In the Filter for year menu, select the year 2010, and click Apply.
  15. For good measure, add Year to the Tooltips menu before clicking Refresh Visual again.
  16. When you hover the mouse over the updated visual, you can see that in 2010, the population of Texas was 25.1 million. That makes a lot more sense. Note that the formatting of the tooltip can be improved in many ways, including alias of fields, and number formatting. We will cover those topics in future articles.
  17. Don’t forget to click Save to preserve all our improvements to the graph. Anytime you have unsaved data, the Save button will be green.
  18. Lastly, before we wrap up, let’s see how easy it is to see the same information through different chart types.
    • In the top left corner of the screen, click Clone to make a copy of the visual.
    • Change the name of the new visual to Superimposed State Populations, and click Save.
    • Remove State from the X shelf (just pick it up and drag it to an empty part of the screen).
      Then, replace it by dragging the field Year onto the X shelf.   You may notice that the year value now becomes sum(year), simply go to year’s Aggregate menu and click on Sum to unselect summing the value of the year.
    • On the Y shelf, remove the ordering: click on the side arrow next to sum(Population), select Order and Top K, and then click on the Descending again to uncheck it.
    • Remove the Year fields from the Filter shelf.
    • In the Visual Type Selector, pick another visual type, Lines.
    • Click Refresh Visual.
  19. Note that the new line visual shows the population increase for each state over time.

That is all for this tutorial on how to create a visual. There are many more visual types that you can explore on your own. Each may have its own unique settings so don’t be afraid to experiment.

Other Getting Started Guides include:
Getting Started 1 – Getting Started with Arcadia Instant
Getting Started 2 – Connections, Data Import, and Data Sets
Getting Started 3 – Working with Dashboards
Getting Started 5 – App Navigation