Sessions 9 and 10 Annotations

Hi all,

The following annotation is regarding the Socio-economic class (SEC) classification I covered in session 9.

• As mentioned in class, the SEC system has changed substantially in this decade. I discussed in class what the old system was like.
• Mr Sharath Srinivas from your class has sent slides that explain the new system (Thanks!). These slides are putup on LMS.
• This Wiki page also explains the change that has happened and the reasons driving that change.
• Pls go through the new SEC system at your leisure. I believe it will help to know about the new SEC in your placement interviews with B2C Marketing firms.
• However, since I haven't discussed it explicitly in class, the new SEC system will not be important from your end-term viewpoint.

**************************

In Session 10, as I wrap-up the course, I champion R as an Analysis (and not just analytics) platform for industry use, try to convince you to include R in your consideration sets, and talk about its learning curve. This post from term 5 also contains some pointers and sources for learning R.

For instance, to demonstrate the wide flexibility and inter0operability of R with so many other packages around, I pick common 'how to' phrases people may encounter in MKTR and google their results on R. Thus, one could ask "How to test hypotheses in R?", "How to sort in R?", "How to do cluster analysis in R?", or "How to do ANOVA in R?" etc.

Some of these results I present below:

Normally, somebody somewhere would have created an R package with basic functions that others would later extend. For running SQL queries in R, look at the first result - a package 'sqldf' comes up. The third result, a PDF also shows up. To my untrained eye, the 5th result too looks promising and so on.

****************************

Here's another instance: say, in your work, you come across a potential opportunity that you want to proof. but it involves text analytics. And your firm isn't currently using any licensed commercial text analysis platform. But you recall MKTR@ISB and turn to R. You google for 'text analytics in R' and here's what you'd find:

The first two results show you the 'tm' package in R. The first link is from 2008 when 'tm' was first introduced as a basic version, its quite enhanced in functionality now. The 'tm' package manual you find in the third search result. The fifth result is a research paper that provides Perl+R programs for a particular function. And so on. Any decent stats grad you hire as an analyst can pickup R in a few weeks. You would have inexpensive access to a great platform.

***************************

Say, you're working on something and you want to merge two datasets. You don;t know how. Just google it. That's how I learned my R in the initial days, getting stuck on something and then figuring out through Google how to do it in R.

Lemme quickly mention Re the first search result: Quick-R is a great resource for do-it-yourself on R. In fact, this blog's format for R code is inspired largely by Quick-R only.

Well, Dassit for now. Over to the last class tomorrow and I hope you would have gained as much from the course as you had expected.

Sudhir