Today it is the world of Data Science. Ample amount of data is available which when utilized at right time in right manner can help to forecast as well as predict in advance any untimely failures/disasters that can cause serious and fatal losses. Out of many, one area where such predictions and Predictive Analytics Software can be of great use is, manufacturing/process industries like power plants, oil and natural gas and many such more.

Predictive Analytics Software not only has components like main stream software-intensive systems, but it also has statistical algorithms, tools, techniques, mathematical components for pattern recognition as well as other techniques like machine learning, artificial intelligence, modelling etc. It also has diverse stakeholders and other connected systems.

Software Architecture is well practiced in main stream software-intensive systems like web, embedded, enterprise applications etc. However, Predictive Analytics being emerging branch, there is a lot of scope for research and enhancement in Software Architecture concepts with respect to such software systems.

With my experience working with Predictive Analytics Software for power plants, in this paper I will talk about following 3 points:

1. Benefits of practicing Software Architecture in Predictive Analytics Software
2. Challenges in using Software Architecture in Predictive Analytics Software
3. Future ahead