Learn Predictive Modeling – Why and How?

Data Analytics, Big Data Analytics, Data & Decision Science and Predictive Modeling are some of the hot topic in the digital world.

Though data analytics and predictive modeling have been prevalent or are being used by the organizations for many decades.

In this blog, focus is to share our view on why predictive modeling is great skills for a better career and provide an overview on what predictive modeling.

Job, Salary and Predictive Modeling

  • Career in Analytics: Due to data everywhere, a focus on analytics and big data analytics has gone up significantly over last few years. Also, renowned publications and industry leaders projecting huge requirements of analytics talents across geographies.
  • Predictive Modelers in Demand: Using predictive model and analytics, relevant patterns are found from the historical and transactional data and used for understanding a business problems and predicting future events. Predictive Modeling skills are great demand and numbers of organizations are looking for talents with Predictive Modeling skills and experience.
  • Benefits organizations: Key differentiations the predictive model output creates are to improve business decisions and impact organization top line (revenue) and bottom line (profit). It is concrete and measurable impact.

What is predictive modeling?

Predictive Modeling or Predictive Model development is a process to find pattern or drivers which can be used for predicting future event.  The process of development is based on data and statistical or machine learning algorithm.

Also a predictive modeling involves understanding of context, data analysis, and discussions with various stakeholders for variety of reasons.

Some of the common examples of Predictive Model applications across industry verticals

Credit Application Scorecard – A credit application scorecard or predictive model is used for assessing credit worthiness of applicants and based on credit score value credit facility (e.g. Credit Card, Personal Loan or Mortgage) is approved.

Response Model – Retailers or Financial Organizations solicit customers for various campaigns. They want to identify customers who have higher probability of responding to a campaign. In the scenario, a response model is built to identify customers who have higher chances of responding.

Analytics Maturity and Predictive Modeling

Next question comes to mind, what are the different types of analytics? Though there is no standard classification, based on our experience some of the analytics projects could be categorized as follow

Reporting and Dashboard: Reports and dashboard to review and understand current state of the business and the processes.

Diagnostic or Ad-hoc Analysis: Data analysis to understand interaction between various factors, root cause analysis and diagnostic analysis to find why certain things happened the way they did.

Predictive Model and Optimization: Build predictive and forecasting models to predict future actions and estimates future values using analytical, statistical and machine learning techniques. And leverage optimization techniques to maximize business outcomes.

Analytics Maturity v1

What skills do you need to deliver Predictive Modeling?

  • Domain Knowledge: Successful deployment of a predictive model requires proper understanding of the business context and challenges. Identification of business problem and formulation of business problem into an analytics problem require understanding of the context and knowledge of which technique (s) could be used.
  • Analytics Tools and Technology: There are quite a few predictive modeling or predictive analytics softwares. Some of the commonly used tools are SAS Enterprise Miner (SAS EM), IBM Analytics (SPSS Modeler), R, Statistica, Salford Analytics (CART, MARS etc), FICO Analytics, Angoss Predictive Analytics and a few others. Base SAS and SAS EG could also be used for developing predictive models.
  • Statistics and Machine Learning Skills: Knowledge of statistical and machine learning techniques can’t be underestimated. Each of the techniques has underlying assumptions and output results & statistics have to be correctly interpreted to ensure right application of predictive modeling.


What are the options to learn Predictive Modeling?

Colleges and Universities

In the recent years many colleges and universities have started courses on Data Analytics, Big data – Hadoop & Technology, Big Data and Analytics, Machine Learning Applications, Predictive Modeling.  These courses have great coverage on theoretical concepts and diversified overview to various statistical & machine learning techniques.   But these courses require extensive time commitment and some of the courses are expensive as well.   Some of the Analytics courses in India are below and in the next blog we will try to cover the list extensively.

  • Business Analytics – IIM Bangalore
  • Business Analytics – IIM Kolkata
  • Certificate Program in Business Analytics – ISB
  • Advanced Certification Program in Business Analytics – IIT Bombay & Huges


Consulting Companies and Training Institutes

Increased focus on analytics for business decisions and also exponential growth analytics talent requirements, number of analytics focused training institutions has been set up in India and across world.  These institutions are focused on providing industry relevant trainings. Typically trainers come from corporate with hands on experience in leveraging statistical & machine learning techniques for solving real life business problems.


A few others institutes also offers data analytics, SAS, R, Big Data and Hadoop based training

Online Courses

A large number of universities provide free online video based data analytics and big data analytics courses for learners.  A few others provide Online Predictive Analytics Degrees and top online free training websites are

In addition to attending courses and learning about data analytics and predictive modeling, one can try reading various blogs and books.  And most important of all, identify problem statement and relevant data sources to make hands dirty, nothing can be better than this.

Some of the useful blogs