Analytics apps or embedded analytics
Posted by Dylan Wan on February 16, 2017
An analytics apps should be an application. We should not see it as a report solution or just a way to view or download data.
A best analytics apps is an apps, and lets people forget about it is an “analytics” apps.
If you keep the end user in mind and drive the design of the apps using user persona and knowing really what users do, the line between analytics apps and other apps should not be very clear.
A teacher wants to see student records, student progress, and need to know who need helps. A teacher wants to improve his or her teaching methods and want to ensure the student learned. They need information but they also need to have ways to comment and provide their inputs. They want to have “watch lists”. They want to enter students into intervention programs. After that, they want to see the performance by intervention programs and by watch lists.
An analytics platform can limit themselves to say “it should be done in transnational apps”, and forget about the requirement, or an analytics platform can take it and address the requirement.
Should this requirement really be addressed by “Embedded Analytics”? Is this really a “Write Back”?
Showing “Tag Cloud” is a nice visualization feature. How about enabling people tagging?
Typically filters are used to narrow down. Once I reach a list that I look for, why I cannot use selection to limit it? Ok, it may be just a IN list or NOT IN list filter. Why not save the ID and a tag in another “generic” object?
The metadata should be very simple: object Name, and primary key columns from that object. The data storage is just a map. The “Write Back” is not really a write back. It should just look like tagging. The internal data storage need to be creative. Do not store them in the relational way. Tags are filter criteria. A list is just a list. If I can do a LIKE, filter by the list members or tags should not be too hard.