A Data Science Central Community
It is now a given that it is extremely difficult, if not impossible, for organizations to survive without 'Business intelligence' (BI) i.e., Analytic applications that help in decision making by providing the right actionable information to the right people in the management at the right time.
For any analytical application to be powerful enough to provide decision support, it has to be both useful and usable. Usability requires that the application be easy to understand, navigate and use; whereas usefulness requires that the analytics provide the right actionable information for decision making.
How do we ensure that these two qualities co-exist?
Usability requires proper data visualization, which in turn requires an understanding of
* the actionable information being presented to the end-user
* the nature of data (i.e., whether it is quantitative or qualitative, whether it is ordinal, nominal, interval or ratio data)
* the best practices to be followed for visualizing data (for instance, the right chart to be used for visualizing a given type of data, etc) &
* the other typical navigational aspects for the application such as menus, navigational hierarchy, data filters, security features (i.e.,type of access privilege required by the various end-users), etc.
The above mentioned aspects must be clearly documented at the time of gathering requirements. These have to be then validated by the end-users at the time of User Acceptance Testing (UAT).
For the analytical application to be useful, the information and the analysis presented should serve to
* reduce the latency in decision-making i.e., enhance the speed of decision making
* help in closing the loop by indicating action to be taken by the management to achieve the organizational objectives within the stipulated period of time
It is to be remembered here that the analytical application itself cannot take action but can only prompt action by indicating the areas of discrepancies. It is also important to understand the role of the end-users accessing the application so that the right kind of information can be provided to them for decision-making.
For ensuring that the application is useful, it is imperative to understand the metrics and dimensions of analysis during requirements gathering. There also has to be a clear mapping of these measures and dimensions with:
(a)the organizational goals to be achieved
(b)the key process areas that would need to be improved for achieving these goals and
(c)the process owners.
Without such clarity in mapping of the requirements, the dimensional modeling exercise, which is a critical pre-requisite for building analytical applications, would go for a toss.
So, in essence, usability quotient and usefulness quotient must be very high for building any powerful analytical application.