This week we are happy to introduce and welcome our guest blogger Nirav Dalal. Nirav is currently working with a large KPO and has close to 10 years of overall experience in the analytics and modeling space. In his role he is responsible for delivery, solution development, business development, research and innovation. He has worked extensively in the areas of sales and marketing analytics, consumer analytics and predictive analytics in the CPG / Retail space. He has a master degree in Statistics and PhD in applied mathematics. Nirav writes about the retail analytics industry covering topics on specific solutions, to generic trends in the industry. In this post Nirav talks about what an analyst’s job really entails.
Getting your hands dirty with data
There are significant number of people employed in the analytics industry who do not have an academic background in quantitative disciplines like Statistics, Mathematics or Econometrics. If you evaluate various solutions applied to solving business problems in the industry, there is always a statistical rigor behind these solutions. There are many folks in the industry, with diverse backgrounds like engineering, environmental sciences, pharmacy etc.who have done exceedingly well in the field of analytics. There are a couple of qualities required for such professionals to succeed in analytics:
In this article let us concentrate on data. This may sound cliché but data is the basic building block of analytics. Without data there is no analytics. Fortunately for analytics, there is abundance of data now. There are companies in the world, whose business is to collect and store data. Digital age has certainly helped and we collect all kinds of data from transaction data in retail stores to medical statistics of new born babies. For people who do not have rigorous training in quantitative discipline like Statistics, they will not possibly have seen large amounts of data. Analytics can involve looking and working with large amounts of data, sometimes millions of observations. This can be overwhelming. Often you would use more than one data source to process the information for business objectives. The problem is not computational in nature, since with powerful computers and servers it has been largely possible to work with huge datasets. I refer to the ability to manually scroll through data and find trends and find reasons behind insights provided to the clients. This is because analytics is all about insights and insights come from looking at data very very closely.
Let us go through an example. Take the case of a simple regression model to know the price elasticity of a product. Price elasticity can be described as the change in sales of a product compared to change in price. A prerequisite to this model would be to look if the price change exists. In order to answer this question, one would need to find the percentage of change in prices of the product for all the stores in the data. A custom code would have to be written to do this. The datasets size could vary from 50000 observations to few lakhs. The analyst would then have to study the price changes to come to a conclusion on the validity of the model.
So if you are looking to join the analytics industry, be ready to look at large volumes of data to justify your results.