HOST: Welcome to the Live Chat for :! We are happy to have you join us. We will be launching our discussion in approximately five minutes. HOST: Welcome, everybody! We have Professor Raghu Iyengar in the room with us today, ready to answer your questions pertaining to the course Raghu Iyengar: hello everyone Vivekanand Sangle: Hello Prof NSB: Hello Prf HOST: Would anybody like to begin today? We'd love to hear your examples of how you are using the tools and techniques from the course, as well as any questions you may have. HOST: How is everybody doing in the course? Chunku Pani: hi HOST: Would anybody like to introduce themselves, and how they are using (or plan to use) the concepts from this course in your work? Chunku Pani: The course is going great NSB: Hi I have to admit I have not been able to progress much on the course I hope to catch up this weekend... Chris Kennedy: Hi, I'm Chris, I currently am in the banking industry and am looking to dive into segmentation analysis. Before looking at company data internally, I was hoping you may be able to point me to some publicly available data so that I can practice the techniques illustrated in the course. Chunku Pani: It would be great if we are shown how the tools work Vivekanand Sangle: I am Vivek. I have not started this yet course. At present focusing on Managing the value of customer relationships. Raghu Iyengar: Hi Chris. Thank you for your question. Chunku Pani: I am working on public data from the US on ailments and people s behavior buying habits etc. Raghu Iyengar: I would suggest that any dataset you find publicly can be used! For instance, CP is suggesting he is working on a public data from the US on ailments. You can use that. Chunku Pani: Have to draw useful insights haven t been able to come up with concrete insights yet
Chunku Pani: Hi sir, I am she Raghu Iyengar: My apologies, CP! Vivekanand Sangle: I am doing a study on how do we use different data like sensor data, GIS data, etc. for defining shopper persona Chunku Pani: Sir, will we be shown how to work on R, Python etc. Vivekanand Sangle: Which of this technique would be of use Raghu Iyengar: CP no. not for this course. This is more of a general course. There are other courses available on edx for R and Python Raghu Iyengar: VS segmentation is very important! Chunku Pani: Sure, sir thanks! Tom: Hi, I have a question about regression. I took a course several years ago where we ran regressions, and if my memory is correct, we converted the independent variables to their natural log before running the regression. Can you tell me the advantages of doing this as opposed to using the raw inputs? Vivekanand Sangle: I know the idea is of doing a segmentation, which is beyond demographic segmentation Raghu Iyengar: Tom good question. Sometimes, some of the variables may be very skewed some very large outliers. So, taking the logarithm can be helpful in taking care of those outliers. Raghu Iyengar: VS Indeed. I would say it will be very helpful to do activity-based segmentation Tom: Thanks. And this should not dramatically affect the output otherwise? Raghu Iyengar: Well, it may. First, the interpretation of the coefficients will be different. This is now the sensitivity to the logarithm. Second, if there are many large values, then taking the log is squeezing them in. Vivekanand Sangle: The question is, there would be data about location, there would be data from the wearables, there would be data from the sensors that track the gaze, data about emotions, etc. I understand the theory. I am ignorant about the practical aspects. Raghu Iyengar: Tom here is what I will do. Do both regressions make predictions and see how they are different. And in particular, for which type of observations. Tom: Got it. Thanks! Chunku Pani: Sir, I am working with an insurance company who wants to gauge the latitude and longitude of various houses near a specific area by using the images of the properties.
Meenakshi: Hello, Raghu, sir. Meenakshi from India. Chunku Pani: Also, whether the houses are in uphill, downhill, near the riverbank, etc. Raghu Iyengar: CP ok, what is the question I can help with. Chunku Pani: We have tools available to identify the same. But what specific insights can I share with them with the limited info that we have? I mean, they have just said we want to know the latitude and longitude Meenakshi: What do you predict as the near future of analytics in the tools side? What would be good to learn like statistics? Raghu Iyengar: What is the business question? That is important to know. Chunku Pani: They are trying to predict how likely these properties are going to get affected by flood Raghu Iyengar: Ok. Do you have some data (not what you collected) but another dataset where you have such information and some floods did happen? Meenakshi: Ok how far do managers need to get into programming/data science? Raghu Iyengar: You can use regression methods to make predictions. For managers, I would say you should be comfortable leading the teams who knows R and Python. So, you may not need to know all the details yourself, but if you are leading a team of programmers, you should understand what they are doing. Meenakshi: I understand Are there any good books that you can recommend for learning statistics as well as for analytics? Raghu Iyengar: I think some other courses from Wharton on Statistics would be very good! Vivekanand Sangle: Did you happen to see my query Raghu Iyengar: VS I may have missed it. Could you please copy it again? I am sorry Meenakshi: Thanks I will check. I had posted a question on the course discussion forum would be nice if you could check that and let me know when you have time questions related to the course content slides Vivekanand Sangle: The question is there would be data about location, there would be data from the wearables, there would be data from the sensors that track the gaze, data about emotions, etc. I understand the theory. My question is how do I acquire analysis and get insights from it. What are the tools and techniques keeping in the voluminous data?
Raghu Iyengar: VS good point. Here is where I would start thinking about Hadoop and storing/analyzing data using clusters HOST: Meenakshi - I think I see the post you mention. Would you like to see any of them mentioned here during the live chat? Nidhi Arora: Hello Raghu...Nidhi from India. I need to know that how can I use the learnings from this course in Indian stock markets industry...from where I will get the relevant data. Vivekanand Sangle: I am a marketing professional. I have heard of Hadoop and am aware of concepts like BIG data, but have not been a user. Meenakshi: They are related to the course slides...so unless everyone has gone through them, it may be distracting Raghu Iyengar: VS I would say then it may be best to find people who can manage that level of data Vivekanand Sangle: So keeping in mind my limitations, are there any tools and techniques from the course that I should focus? Chunku Pani: Hi Prof, what is a general step by step approach to a data analytics project? Meenakshi: I think you can respond publicly on the thread...so everyone can see at leisure Chunku Pani: Or it depends on the business problem Raghu Iyengar: Nidhi from this class, please see the session on regression analysis. One can easily imagine using some of those techniques for stock prices. CP Indeed. It depends on the problem. Incidentally, problem is the most important thing! Nidhi Arora: What is the time frame required for identifying the stock prices data (past data) and how do I measure the error in that? Meenakshi: You can respond on that thread it can be distracting here. Raghu Iyengar: NA for that, you will have to check some finance texts. I can talk about the methods but the specifics depend on the problem Chunku Pani: Sir, for the case you have taken up in Unit 2, which tools have you used for regression? Raghu Iyengar: I used simple Excel. You can use JMP, R, and other software too. Chunku Pani: Does the usage of the tool depend on the volume of data? Raghu Iyengar: Indeed. Excel may not be best for very large volume of data.
Chunku Pani: Sure, sir. Vivekanand Sangle: Keeping in mind that AI is going to be future, to what extent learnings from this course would be useful? More importantly, what kind of quantitative skills should professionals like me need to build? Raghu Iyengar: VS many of the inputs in AI are basically regression models. In this course, I have talked about the basic applications of regression. The next step is to learn about machine learning tools such as neural nets Kruthika: Hello, sir. Kruthika here from India. Thank you for this opportunity. Two questions - 1) I want to cull out insights for an NPD using social media. The outcome for me here is getting actionable insights. Now am confused as to how to measure the ROI here could you please throw some light? and 2) How do you see the Digital marketing landscape, esp. the 'observation' methods evolve in the background of increased focus on cyber security? Would we not be blacked out by our TG? Vivekanand Sangle: Do you recommend a follow-up course that would be useful for me? Raghu Iyengar: K Good point. There is a lot of noise in social media. The key thing is that ROI can be hard to measure in $ terms unless you are doing a careful A B test Kruthika: Sure, sir. So is there any other way to convince the CFO with some quantifiable result? Chunku Pani: I am just in Unit 2, so don't know whether it's already there in the course videos or not...but would it not be good if we are shown some real-life cases where data have solved some business problem? Raghu Iyengar: I think what is important to convey I that you may have to do a careful analysis. Please see this: http://www.drvkumar.com/creating-a-measurable-social-media-marketingstrategy-for-hokey-pokey-increasing-the-value-and-roi-of-intangibles-tangibles-2/ Kruthika: Thank you, professor! Chunku Pani: Like today, I was reading Disney is using lot of analytics/predictive modeling for a seamless consumer experience. Similar to these lines? Raghu Iyengar: CP thank you, we will certainly consider this in the future. Chunku Pani: Thanks, professor. Sir, I have heard it s more difficult to convince a CIO than a CEO when it comes to digital transformation. Vivekanand Sangle: Are you considering analytics for seamle customer engagement? Because my question is related to that Chunku Pani: Can you throw some light on how to address a CIO s concern?
Vivekanand Sangle: The idea is to drive long term customer value and customer referral value using multiple data Raghu Iyengar: CP usually the problem is that CIO may have some legacy links to one or the other type of technology so it can become hard. VS I am not sure what you mean? Vivekanand Sangle: You had said that you are considering analytics for seamless customer experience Raghu Iyengar: VS Indeed. One has to make sure you are collecting good customer data all the way Vivekanand Sangle: So, my question is that are you considering covering customer analytics, which are beyond transactional and demographic data? Raghu Iyengar: One should, but it becomes hard to get good data. Vivekanand Sangle: I am given to understand companies are now looking at decoding the consumer DNA using behavioral data, location data, etc., etc. Raghu Iyengar: Definitely! You should read this book: https://www.amazon.com/tap- Unlocking-Mobile-Economy-Press/dp/0262036274 Chunku Pani: https://www.forbes.com/sites/bernardmarr/2017/08/24/disney-uses-big-dataiot-and-machine-learning-to-boost-customer-experience/?s=trending#5133fcad1365 Vivekanand Sangle: Thanks. Raghu Iyengar: CP thank you. I had read that article and it s fascinating! Chunku Pani: Thanks, sir. HOST: Does anybody else have questions, articles, examples to share? We have just a few minutes to cover any further topics. Chunku Pani: Is this the only chat session that we will have? HOST: Yes, this is the only live chat. Although, we encourage you to continue the conversation in the Discussion Forum. Chunku Pani: Sure, Erik. HOST: If there are no further questions, we want to offer our thanks to everyone for coming! And thank you, Prof. Iyengar, for sharing your time today. Raghu Iyengar: Thank you, everyone! Chunku Pani: Thanks, professor. Thanks, Erik and everyone.
HOST: Have a great day! (or night, whenever you might be located.) Kruthika: Sir, whenever I see the major disruptive market entries, I am increasingly prone to think that they would not have relied on historical data or need-gaps of customers to a great extent; partly because there wouldn't have been any. What worked seems to be a strong intuition to work behind 'that aha' idea is there any credence to my thinking? Ok...time out, I guess Thanks anyway, team! HOST: Thanks, Kruthika. We ll send it to Prof. Iyengar to ensure that you get an answer. -END-