It is amazing how much data is available to just about every business these days. Data flows from every which way, leaving us to determine how to use it. While in the past, we struggled to gather the data that was needed and to secure enough hardware to store it, our challenge now is to figure out how to translate data findings and communicate them as clearly as possible.
This is where embedded visual analytics excites me the most. It allows us to communicate only the information that is needed at hand rather than attempt to explain the entire universe in one view. With embedded visual analytics, we can create crystal clear reports within a specific context, a particular product segment within a sales branch, or automotive issues experienced for a specific make and model based on geography. We can create reports that show users exactly what they need to see without the confusion of information overload.
Consider, for example, the vast amount of data that is available with regards to baseball statistics. Enormous amounts of data are being collected around batting averages, pitches, and fielding. As you can imagine, with the number of teams and players to have played the game, there are billions of records of data available. To filter on this data manually can be quite an undertaking.
In-context reporting with embedded analytics will provide you with faster, more accurate answers.
So how do we go about building in-context, embedded visual analytics? In short, we start by developing a high-level, all-encompassing report as our main template. We can then apply various filters and other features for each of the context-specific reports. Finally, we embed the report and pass it on to the business users. Let’s look at each of these steps in more detail:
1. Create high-level report
Yes, the goal here is to create an in-context report that does not bog the user down with information unnecessary to the task. However, we must start with the full universe of our data before determining how to zoom in to the information that is needed. The beauty here is that we can use the same high-level report as the starting point for reports pertaining to different contexts.
In our baseball example, we might create one overlying report that shows the players who lead the league in batting average, runs batted in, or home runs. When users visit the page for a specific team, however, this report would then be filtered for only players on that team.
2. Apply filters and features
Once the main report has been built, we can then create the desired in-context reports. The obvious example is that we can create player reports specific to each team in the league. However, we can also expand this into more obscure contexts, such as players whose fans are mostly in a certain country, or players who are experiencing their first year in the league. When it comes to the contexts that can be applied to embedded visual analytics, the possibilities are endless.
The same can be said for embedded reports for business applications. We might create an overlying report with all of the consumer insight data that we have available, and then spin off various in-context reports pertaining to specific cities or demographic groups to help us understand how sales vary by those segments.
3. Embed the report
Embedded reports empower the user.
If you were browsing your favorite baseball website often, you likely would become frustrated if you needed to manually filter down to your favorite team every time you wanted to view the team’s leaders. Instead, you likely can go to go to that team’s page to view its statistics. Likewise, a sales manager should not need to manually filter sales and consumer insight data every time he wants to view reports for a specific branch. In-context reports empower the user to quickly get to the information that is needed without the frustration of manually filtering or manipulating the data.
For another example of how to create in-context reports for embedded visual analytics, check out this video, which shows the creation of an embedded report pertaining to vehicle complaints, which is then filtered by make, model, and even vehicle crashes.