Click on the titles to read the articles –



Marketing Campaign Management || Decision Trees || Neural Networks || Model Selection || Linear Regression || Profitability Estimation 


Customer Segmentation || Cluster Analysis || K-Mean Cluster || Cluster Optimization || Customer Segment wise Business Strategy


Credit Risk Modeling || Risk Scorecards || Credit Underwriting || Logistic Regression || Model Estimation


Sales Forecasting || Time Series Modeling || ARIMA Models || Real Influence of Marketing on Sales || Add Predictors to Time Series 

Regression Analysis Case Study Example 1

Estimate price using regression analysis || Principal Component Analysis || Ridge Regression and Machine Learning


Risk Management  ||  Data Visualization || Logistic Regression with WoE & IV  || Reject Inference


Business Process Optimization || Agent and Customer Segmentation || Linear and Integer Programming || Real Time Analytics


Best Practices in Data Science & Analytics || Software || Project Lifecycle || Quantify Business Benefits

A/B Testing || Maximize Returns on Marketing  Investment|| Bayesian and Reinforcement Learning || Thompson Sampling


Application of Big Data Analytics || Real Time Analytics || Business Benefits


Application of Bayesian  Statistics || Business Usage || Bayesian in Big Data Environment


Business Analytics to Improve P&L || Business Thinking || Problem and Data Discovery for Profitability


EHR / EMR Data Analytics || Proactive Diseases Management || Wearable Devise Analytics || Payer & Provider Analytics


Books to Learn Skills for Data Science || Download Free PDFs || Self-Learning Strategies


Problems and Challenges for Data Science || Work with Data Science Community || Hands-on Experience

Analytics and Date Science Career

Career in Data Science & Analytics || Job Interviews || Best Practices to Excel in Your Career


Conversation with Industry Experts || Learn New Ideas || YOU CANalytics Event Participation

Learn Python and R

Free Books to Learn R & Python || Machine Learning  || Data Science  || Artificial Intelligence


Hands-on Experience of Data Science || Tools and Software for Data Science || Data Exploration and Analysis


Novel Ideas in Data Science and Analytics || Emerging Trends ||  Business Process Integration with Data Science

Machine Learning Institution || Optimization  ||  Gradient Descent and Regularization   || Cross-Validation 

Artificial Intelligence || Deep Learning ||  Neural Networks Simplified


70 thoughts on “Blog-Navigation

  1. Roopam, I happened to accidentally discover your blog while looking for something else. There’s just one word to describe it – PHENOMENAL. I haven’t been able to get off it for the past 2 days and hope to complete each and every word you’ve written. It’s taking me this long, because of a busy schedule, else I would have stayed glued completely. I have till date not found a single blog which makes learning Analytics so easy to follow and implement.

    Just one final request – never stop writing – we (read I) need you to continue this forever 🙂 Absolutely mind-blowing stuff and the clarity and ease with which you are able to relate it to examples in my daily work life. Just awesome.

      • Can you suggest me some links for customer acquisition channel strategies in-particularly banking(credit card,mortgages) and analytics help to improve channel strategies

        • Identification of right channel strategy is one of the key questions to improving sales for an organization. To begin with, you ask a simple question

          1) How are different channels for selling credit card (channel partners, web-site landing pages, branches etc.) performing?
          – Create right metric to do apples to apples compare these channels

          2) A more important question is: how to improve performance for a particular channel?
          – This will require some serious design of experiments followed by multivariate analysis

          3) Finally, you do your cost benefit analysis to identify a right strategy for your bank / financial company.

  2. Roopam, your articles are phenomemal… Extremely informative and easy to follow. Your analogies to day-to-day situations is very amusing.


  3. Can you suggest me a decently easy book on Bayesian Inference to start with( for data analytics)? Something with real life/relevant examples (similar to the kind of stuff u have used to explain). Cheers!

    • Actually, I am planning to write a post about books to self learn Bayesian inference really soon. In the mean time you may want to try Doing Bayesian Data Analysis: A Tutorial with R and BUGS by John K. Kruschke. This is a really good book and the author has a good sense of humor. You could actually find the pre-released draft of this book on the internet. The following link has the first six chapters of the book – Link

  4. Hi Roopam,

    I am taking my first baby steps towards a career in analytics. As I have started reading your articles, it has filled me with joy. It would be interesting to correspond and get in touch.

    I currently work with an organization in the midwest of the USA. 😀

    My self training on analytics owes a lot to you and also to the folks in Analytics Vihya.

    Thank you,

  5. I was wondering if you might be able to explain some cases in market mix modeling by using example datasets. It would be great.

    Take care,

  6. Hi Roopam
    Nice repository…Its really fits in my line of thinking – ‘Keep it simple’…All the best and please keep writing

  7. Hi Roopam

    If you could recall we were in contact when you were in axis bank for some business related queries.
    Just wanted to compliment on your great work on the blogs here. I really enjoyed your blogs on customer segmentation .

    Could you recommend me something on personalized offers for retail customers ?

    Thanks and regards

  8. Dear Roopam,

    Thank you very much for your extremely valuable posts. If you could provide some material for SME scorecard development, it would be so great. I have a question on SME credit score: why SME credit score doesn’t have behavior scoring model, while Retail scorecard has both application and behavior scorecard.

    thank you very much!!!

    • Thanks Thare,

      SME scorecards are conceptually similar to retail lending scorecards because usually SME loans portfolios, unlike large corporate loans, are also reasonably high volume portfolios like mortgages etc. However, the fundamental difference between these scorecards is the same as difference between a small company, and an individual. Hence, your input variables for SME scorecards will be quite different.

      There is absolutely no reason why you can’t have behavioural or collection scorecards for SME loans. One reason sometimes banks don’t opt for behavioural scorecards for SME is because of their relationship network with SME clients which they believe is a better indicator of solvency of the SME. However, as the diversity and volume of SME portfolio grows behavioural scorecards are inevitable. Also, data quality is one big challenge with SME scorecards.

  9. Dear Roopam,

    Landed to your blog for the first time today, and must say that you have a great writing skill. Loved the way you have detailed on every topic. I have been reading it for a couple of hours and am glued. Will surely read each of them and try using it at work. Thanks for your posts. Keep sharing your thoughts.


  10. Hi Roopam,

    Thanks for this great blog. Better than any other blog or book i have come across on statistics. Do you have any post on sampling (sample size etc) strategy and approach for designing tests in banking/credit cards.

    • Hi Girish,

      Thanks for the kind words. Let me share a piece from a previous article on sampling strategy, hope you will find it useful:

      A few years ago, I did a daylong workshop on Statistical Inference for a large German shipping & cargo company in Mumbai. At the time of Q&A session the Vice President of operations asked a tricky question, what is a good sample size to achieve good precision? He was looking for a one-size-fit-all answer and I wish it were that simple. The sample size depends on the degree of similarity or homogeneity of the population in question. For example, what do you think is a good sample size to answer the following two questions?

      1. What is the salinity of the Pacific Ocean?
      2. Is there another planet with intelligent life in the Universe?

      In terms of population size, number of drops in the ocean and planets in the Universe is similar. A couple of drops of water are enough to answer the first question since the salinity of oceans is fairly constant. On the other hand, second question is a black swan problem. You may need to visit every single planet to rule our possibility of intelligent form of life.

      For credit scorecard development and other classification problems, the accepted rule of thumb for sample size is at least 1000 records of both good and bad loans. There is no reason why you cannot built a scorecard with a smaller sample size (say 500 records). However, the analyst needs to be cautious in doing so because a higher degree of randomness creeps in a small data sample. Additionally, it is also advisable to keep the sample window as short as possible i.e. a financial quarter or two while scorecard development. Further, the sample is divided into two pieces – usually, 70 % for development and remaining for validation sample.

      Link to the above article

  11. If you were to develop & track analytics for a training organization what would be the key ones to focus on?

    • Hi Bryan, sorry for delay in response to you question. I missed your comment.

      These are some of the steps I would follow to track analytics for a training organization:

      1) Ask the relevant questions that are most important for the growth of a training organization such as
      – Training Impact: How are participants getting better after the training?
      – Trainer Impact: Is a particular trainer better at explaining certain questions?
      – Word-of-Mouth Metric: are the participant talking about your training on social media and other channels?
      – Marketing Metric: Which channel is bringing most business for the organization?

      These are of course a short list of questions. I am sure you will have a much more detailed list.

      2) Create scientific metrics to quantify results to answer the above questions. For instance, the first question about participants growth can be tracked through their pre and post training performance.

      3) Create experiments (design-of-experiments & A/B testing) to track and analyse the above mentioned metrics along with predictor variables such as education background of participants, years of experience etc.

      4) Analyse the above results to improve the key performance metrics.

  12. Hi Roopam,

    I think you are doing a great job (most people here agree with me). And like Kisalay said, never stop writing. Please write some basic analytics too (if you think that’s appropriate.


  13. Hi Roopam,

    I got ur blog details from the linkedin profile. The way you have described some of the concepts is really awesome. I am fascinated by the depth and simplicity of each topic and ur style of writing. I really want to meet you. Are you part of any meetup group or any related analytics club in Mumbai. Please do share.

    Amit Kumar

    • Hi Amit,

      Sorry for the delayed response was really tied up lately. I am not a part of any meet-up group but we could meet up, please drop a message in the contact section, and I will share my coordinates.


  14. Hi,

    I also happened upon your website and want to say “kudos” for this work and site. I love the way you categorize the cases and the insights into scientific research you provide. I love this area of customer analytics and providing only what people want or need at the time they want or need it. I enjoy your articles, keep on writing!

    Best regards,

  15. This blog is turning out to be a survival kit.Too too good articles.The best part is explaining the complex stuff in ease.Also the depth in which you cover is absolutely phenomenal.Was thinking if you are planing something on survival analysis with big data sets.

    Thanks a ton,

  16. Hi Roopam,

    Everything is analytics explained in a very simple way. Keep up the good work. Is there any way of following or subscribing to your blog such that we could get intimations of any new blog you post?

    Piyush K

  17. HI Roopam,

    you have expalined lot of things in the easiest way. Great. Thanks.

    Just one suggestion – kindly add a search text box in your website. it will be realy helpful for people like me.


  18. You are one of the most creative and innovative writer I have seen who demystifies the nightmares of Analytics. BRAVO on the good work and I hope I can someday contribute to your work with such exceptional clarity and fun examples.

    PS- Your interesting examples of Sherlock, Batman etc are sometimes distracting 😛 :P..


    Your new fan
    Amit Jain

    • Thanks, Amit, for such kind words.

      Many times, as I have learned with experience, controlled distraction is not a bad thing. Since, it helps creativity, which in my opinion, is at the core of analytics and problem solving. All the best – look forward to reading your posts.

  19. Hello Roopam

    I loved your articles.I find it very easy to understand rather than going through books on analytic and data mining. I have one request though . Could you please add any case studies on Supply Chain,Operation and Procurement analytic.

    • YOU CANalytics is an individual effort so far. At some point, hopefully soon, I would like to build a team to make this a more organized and dedicated endeavor but as for now, it’s all me.

  20. Hi Roopam,

    Thanks for informative details on different books of R language with reviews.

    It help me to find out various topic learning from different books. I was searching on many places and found here on single click.

    I am new bee in Analytics and have no prior experience.Just started learning basic R from free courses available from internet so is it good start to self learn R from different e-books and moving further ?

    Please advise.

    Thanks & cheers

    • It totally depends on your preferred style of learning. I think, online courses are a good place to start, however, books can help you become hands-on with coding. Online courses can’t provide the level of details you could get from books. I prefer using all the mediums at the same time. All the best!

    • The way you have explained the concepts relating it with day to day activities is making me look at every activity with new perspective!

  21. Amazing Work Roopam… i generally don’t leave any comment, but the way you explained the things forced me to write this…God bless you roopam 🙂

  22. I happen to run into your website by accident when googling for “marketing analytics case study” and I must say that the contents are phenomenal. Have you thought about compiling your blogs and making it a book?

  23. Thank you for blogging about this topic. I work in consulting but am not a modeler so I really appreciate your plain-English style of explaining concepts.
    Often analytics/statistical blogs miss hard on communicating the technical pieces but you do a stellar job.
    You’ve won yourself a subscriber. Please keep up the great content!

  24. Simply great…I never read such a nice tutorial among the complicated books with X^Y^>>><<<<<++___^^

  25. Hi Rupam,

    Again no words to describe knowledge you posses and they way you share in open forum. Hats off!!

    I am working for Insurance company in India and would like to understand more about analytics on
    – how to analyse and identify social media leads ( google ad words / Facebook etc)
    – Dashboard analytics on revenue/ cost on region wise to take better decissions.

    please let me know if you have any articles with the same context

  26. magnificent put up, very informative. I wonder why the opposite specialists of this sector don’t notice this.
    You must continue your writing. I’m confident, you’ve a
    huge readers’ base already!

  27. I am genuinely happy to glance at this weblog posts
    which contains tons of valuable information, thanks for providing
    these kinds of statistics.

  28. Hi Roopam,

    I am trying to do ECL forecasting. Basically, we already have ECL model for PD, EAD and LGD which we run month on month to calculate the expected loss but now i want to go one step further to even forecast my ECL numbers month on month for next 12 months and for that i am trying to create forecasting model for PD, EAD and LGD. Now the challenge is it is basically a forecast of forecast and don’t know how to go about it? if you can suggest some ideas that will be great.


    • ECL essentially depends on factors internal to the banks and economy/surrounding factors. I am assuming your activities of forecasting ECL are driven by the COVID-19 pandemic. In this case your existing ECL model will not be very useful since your PD and LGD models are mostly not equipped for the unprecedented challenges this pandemic has created. I would suggest you use the bucketed segment from your PD and LGD models and track the influence of economic factors on your segmented portfolio. You could use exogenous variables ARIMA or neural networks to model ECL for these segments. A 12-month forecast in such a volatile time will be quite tricky so identify the influence of the external factors (i.e. job losses, sectorial GDP, post moratorium bureau reports etc.) on your segments and track these factors to make the forecasts. You ECL forecast will be driven by the forecasts of these external factors. I must say you will have to reevaluate your forecast on a monthly basis with the performance of your portfolio and change in these external factors.


Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.