ArcGIS Insights & Dashboards
| Topic |
|---|
| Overview |
| Insights vs AGOL Dashboards |
| Where to go from here |
| Next: Demonstration of ArcGIS Insights |
Overview
Data visualization has evolved tremendously in the last several years. For decades now, software applications like Excel have enabled common users to create plots, graphs, and charts of data, allowing for much more effective communication of results. Below is an example of a very rich visualization from the GapMinder project. Note how many dimensions of data are shown for each country and how they are shown:
- GDP - long the X axis
- Life Expentancy - across the Y axis
- Population - as size of the symbol
- Region - as color of the symbol
They could even add another categorization of value by altering the shape of each symbol.
A good visualization allows the user to glean multiple insights from the same dataset, as this one certainly does.
More recently, however, data visualization to another significant leap in the form of data dashboards. Beyond nifty arrangement and symbolization of data points, these dashboards allow interaction with the data: sliders, selectors, filters, connectors. These combine to allow dashboard users to dig deeper into the data and gain greater insight.
The folks at GapMinder have an excellent dashboard version of the same dataset found at this link. Open it and have a look!
- Note that you can view change over time by activating the time slider at the bottom of the visual.
- You can filter the countries shown based on categorization or manually by name
- You can re-color the countries by different categories or values (e.g. Child Mortaility)
- You can select points to gain more information about them.
Clearly, dashboards engage the user at a different level and provide a vast amount of information at much simpler and more digestible level.
However, while dashboards are easy on the user, they take time, effort, and much thought to produce. Behind every dashboard are datasets that have been carefully wrangled and tidied so that the various queries in a dashboard can be executed cleanly and efficiently. Additionally, the dashboards have to be crafted using code and/or software that provides a logical and intuitive interface to the data, and allows the interface to be displayed and shared to a wide audience.
That’s the bad news (for us). The good news is that the code techniques and software platforms to do this are getting much more user friendly. Long gone are the days where you’d have to be an expert in SQL, HTML, and JavaScript to get the simplest interactive widget on a web page. Now we have applications like MS Power BI, Tableau, R-Shiny.io, ESRI Insights, and ArcGIS Online that allows us to create and share dashboards with relative ease (“relative” being the key word here).
In this set of lessons, we do a quick exploration of ESRI Insights and ArcGIS Online Dashboards, focusing on the technologies that allow us to leverage our existing knowledge of ESRI products and geospatial analysis.
Insights vs AGOL Dashboards
Why does ESRI have two products that create dashboards? Well, actually, it has more than two, as Dashboards are just one offering among a suite of Web Apps that are part of the the ArcGIS Online ecosystem.
ESRI’s Web Apps
Open ArcGIS Online: https://dukeuniv.maps.arcgis.com
In the upper left,click on the 3x3 set of dots to the right of your account name.
There you’ll see all the apps within the AGOL ecosystem, among these you’ll see a few that are related to dashboards
Note that this is a rapidly evolving area among ESRI’s offerings; options may appear and disappear at any time!
♦ Insights
- Insights allows for quick development of interactive visualizations of spatial and non-spatial data.
- The emphasis is on speed (and somewhat on reproducibility), not on the overall aesthetics of the output.
- Unlike AGOL Dashboards, you can do some wrangling, editing, and even analysis within Insights, making it something more of a hybrid between ArcGIS Pro (in particular its data engineering capabilities) and a pure dashboard application.
- You can still make your output professional looking and share it, but to make your dashboard really pretty, you’d likely be better off with Dashboards or other AGOL Web Apps. The advantage with Insights is that you’ll typically build your interface much faster.
♦ Dashboards
- These allow you to “quickly” develop interactive dashboards from existing web maps, not relative to Insights, but relative to other Web Apps in the AGOL ecosystem.
- The overall functionality is limited (though still quite powerful), which is the trade off for easy of set up.
♦ Other ESRI options…
Instant Apps
- Instant apps is something of a “wizard” driven interface that guides you toward developing a shareable visualization of your data.
- It’s perhaps a useful place to start, but you may be limited by what’s shown to you unless you actively explore.
Experience Builder
- This is an interface to a number of templates used to develop more powerful dashboards
- Within each template, you have access to a wide variety of widgets that are linked to AGOL assets
- These require more time to construct, but you have much more overall control of what you want to create.
Which to chose?
We are at an exciting – and turbulent – time with respect to creating dashboards. The technology is still young, and nobody quite knows what will bubble to the top as the “best” platform for developing a dashboard. It will likely be some hybrid of these choices. So for now, it’s less important to think you making the correct choice for what you want to do and more important to explore options and learn the key elements shared.
Where to go from here
As mentioned above, with platforms continuing to evolve so rapidly, the best approach to learning dashboards is to try a few options and make note of what’s common about all dashboards. Thus for this mini-section, you will run through two tutorials: one that I have constructed on Insights, and an ESRI tutorial on dashboards.
While doing these exercises, make note of the following commonalities among the two platforms:
- What format do the data have to take to be successfully integrated into a dashboard? More specifically, what prep work (wrangling, grouping, transforming, etc.) might you have to do to get a dataset you’ve worked with into a format that can be used in a dashboard?
- What are the visualization options for your data as determined by the platform and how easy are these visualizations to create? For example, can you easily create bar plots, box plots, etc.? Maps with various symbolizations?
- Same as above, but for data queries and filters? How can you enable the user of your dashboard to select subsets of data, either by query or by graphic selection?
- Related to the last two, how are various components in your dashboard linked such that a selection in one widget triggers a change in another widget.
- How do you provide access to your dashboard to others?
These are the basic elements in all dashboards, and the platforms mostly provide different levels of ease and customization of these elements.