- Create Hypothesis: One of the key steps in modeling process is Hypothesis Testing. Unfortunately, I have seen many analysts considering this most important step as a ritual to be performed. I was and I still am particular of writing down my hypothesis. Without complicating me in jargons of Null Hypothesis / Alternate Hypothesis, I simply consider hypothesis as a subjective opinion of what you feel the trend / pattern would be.
- Validate Hypothesis: Next step is obviously validate the hypothesis using statistical technique and more important by creating a visualization (visual depiction) of the data for your hypothesis; simply put the bad-rate / good-rate graph or the cross-tabs that is often created in most of the modeling process.
- Interpret Hypothesis: Interpret the output of visualization; make your own take-away from the charts and graphs. Do not leave this job to others, at least in initial years of your career. Interpretation will build your knowledge; some of these hypotheses will get repeatedly validated and they will become part of your memory stack as wisdom (domain knowledge).
Business analytics is the art of using the logical reasoning in day to day business. Decisions are made on the outcomes of these logics. To support the logic we need science and thus data science. Fire and wheel brought a revolution some centuries back. Data is the next big revolution and big data is the hot buzzword of today. When it comes to data you speak of Statistics and why not. Statistics is the science of interpreting the data and predicting the outcome through different design of experiments. The quest has always been around How? Why? and What?. Believe me these are the most powerful words ever invented. Analytics is just the story telling but no lullaby. Algorithms/logics/techniques all come out of statistics. The notion of making human capital decisions based on statistical data may sound like a crazy, futuristic concept with a science-fiction twist. In reality, business analytics is a phenomenon that happens continuously within our society, well radiated in this day and age of internet and social media. Just think about all the ads we see every day on our web browser: We see certain messages because they are linked to our data and to statistic generated by our “user experience” (i.e. what kind of things we usually search on google, what music do we listen to, where do we live, what language do we speak…).HR Analytics is a solid reality as businesses shift paradigm from a traditional corporate hierarchy to a work environment in which productivity is stimulated (and rewarded) by taking into account individual characteristics, as well as emphasizing on the value of teamwork Business analytics provided the key measurement at same/different levels/platforms. What makes a difference between good better and best? What makes a craftsman, a good craftsman? What makes a pot-maker, a good pot-maker? What makes a painter, a good painter? Anyone in their respective profession, making use of the similar tools & techniques can create a differentiation by his art, dexterity, skill in using the respective tools & techniques. If modeling is an art, then the question is “How do I develop this art?” The answer lies in modeling process itself. Hypothesis Testing.