Building native applications for mobile devices is like building a ‘Client-Server’ application which was popularized in the 90’s and 00’s with the explosion of personal computers. While this trend leveraged the fact that an application can use the computing power of the client device and provide better user experience and performance. However, from a maintenance perspective this was a nightmare. I still remember the “DLL Compatibility” issues driving me and my team crazy when we were deploying these applications in an enterprise environment and having to update/upgrade the application frequently. It generated a lot of call volume to the support teams. Combine that with the lack of good collaboration tools and you have a perfect storm whenever there is a new release.
The reason I bring this up is that the web has changed this whole ‘Client-Server’ paradigm and drove the application development model to adopt web-based architectures. With respect to Mobile Application Development we are looking at another such paradigm shift where there should be more and more applications built using ‘HTML5’ based app development for mobile devices. Just as in the case of Client-Server apps, ‘Native Apps’ for mobile devices provide a lot of power for developers to build apps that provide better user experience and performance, but it is a maintenance nightmare and also tend to increase the deployment cycles. A ‘HTML5’ application that follows the “80-20” rule and uses 20% native code will provide greater flexibility by delivering almost all the advantages that a ‘Native App’ would. Plus, it also fits into the “Build Once Deploy Anywhere” model where the app can be deployed to multiple mobile devices (iOS, Andriod, Blackberry, Windows Mobile etc.) with minimal changes (to the 20%) of the native code (all the HTML5 code doesn’t need any platform specific changes).
This approach is being recommended by many analysts, thought leaders and vendors. It would be interesting to see how many developers adopt this approach and accelerate the shift in Mobile Application Development paradigm.
Wikipedia defines Predictive Analytics, as “Predictive analytics encompasses a variety of statistical techniques from modeling, machine learning, data mining and game theory that analyze current and historical facts to make predictions about future events.” While not all of the techniques stated above are required, a lot of data mining and statistical analysis need to be performed on historical data before one can predict the future trends and outcomes. The accuracy of the prediction depends on the variables and assumptions considered and will be the key to making accurate predictions of risks and opportunities.
Case in point, the volume of cases that come in for a product safety case processing organization varies depending on many factors. The variance could be due to factors like Seasonality of Adverse Events, A news item discussing potential side effects of a product, A blog post by a physician or some influential group or organization so on and so forth. With the ever increasing cost pressures on life sciences organizations, it is very difficult to plan for peak volumes while there will be additional unused capacity during the troughs. This is the kind of situation where any organization can use some predictability so they can plan the capacity within a reasonable deviation thus normalizing the peaks and troughs.
Imagine an business intelligence solution that can mine the historical case volumes and the corresponding capacity in conjunction with the process efficiency and be able to predict the future capacity requirements. Add the ability to evaluate some ‘What if” scenarios where one can change variable like case volumes and be able to predict the capacity requirements. While this may sound like a “Holy Grail”, it is possible with some of the sophisticated tools available. A screenshot of one such solution below:
Do you have a need for such solution in your organization? Have you built a predictive analytics solution for other purposes? Please share your feedback and inputs.