Quarter 10 - Week 9

My classmates are really restless... it's two weeks to freedom for most of them, and the tension in the air is palpable. Some of the guys are tearing out whatever hair they have left, despite everyone telling them to chill... only two weeks left. Some of the others are wear blank expressions, head in one hand, the day's case study in the other, and people come by saying "Hold on man, just another two weeks!". The combination of rushing around for project work, and knowing freedom's this close is tearing our people apart. Stay the course, people... just hold on... the end is nigh.

Business Data Mining & Decision Models
We got a little deeper this week, with sessions on text mining and web mining. The first was taken by a guest speaker, who discussed what his company does, and how they use their mined information (collected from various sources like blogs, social networks etc.) to identify what people are saying about a brand or organization. Now text mining isn't very simple, as the word 'Care' has a very different meaning from 'Dont care'. So it's the string of words that actually has more meaning than a single word itself. A lot more hullabaloo about how the summary is being done, and after some examples of actual customer projects where they've worked, it 'kinda' got clear as to how text mining is done.

The second session, being about web mining, was easier to understand. Easier since the prof talks about stuff like how you move from one page in the site to the other, how you understand when the usage on the site peaks, what are users more interested in etc. All the more fun when you consider the example he takes - of a matrimonial website. So he goes on to explain what percentage of people actually make a purchase/payment on the site, and what are the most popular pages etc. One of the strange findings was that traffic peaked around 4 AM in the morning, and the prof wonders why the vast majority is logging in that early in the morning in the first place. Then comes enlightenment - the time's according to US time, and there's an audible signal of understanding across the class. Fun stuff!

Strategic Thinking and Decision Making
We're still on the emotion bit... why we decide what we do. Today started off with understanding how our brain functions, and how we make decisions. Apparently, two parts of our brain help to make the decision - one is more dependent on our primal instincts, the other tends to reason. So, the fact that people tend to wait instead of jumping to conclusions show that they rely on the second part, as against people who react immediately and impulsively. The prof goes on to show us some images, and asks us what sort of emotions they incite in us... and based on that, he claims why certain emotions are more likely to cause larger effects than others.

We then go on to the concept of ethics, and personally, it was a rehash of what we studied in EBM in Quarter 8, so just a few new nuggets there. Oddly chilled out class this time, it's getting frighteningly calm. We know what that means...

Quarter 10 - Week 8

Last week was a brilliant break from the typical rigour (*cough*) that we face on the weekends, on account of Diwali. Now, any sane person would think that you have a holiday so that you can relax... but obviously that sane person hasn't done the PGSEM, much less prepared for a project that was due to be given in three weeks! Three weeks? That's a lot of time, you'd say. And I'd agree, if it wasn't for the fact that WE'RE NOT GETTING ANYWHERE!! WE HAVE THE DATA AND THE DAMN THING ISN'T ALLOWING US TO MINE IN ANY REASONABLE WAY! WHO THE HELL INVENTED DATA MINING!! It's a good thing I'm taking this quite well, and not over-reacting.

Business Data Mining & Decision Models

We had the research assistant take our class, and he typically does a good job. But he sort of over-estimated our concentration this time around. Our man is trying to take us through a practical demonstration of how people choose what liquid detergent to buy, based on brand, price, quantity etc. And how the effect of price can affect purchase decisions.

With all sorts of magical flourishes, he stands at the front of the class and with eyes glistening with the promise of the future, and no mic, explains some fundas which I couldn't hear. But they must have been good, as the front benchers weren't looking all around. Some of the guys out at the back are straining their ears, hunched forward, staring pointedly in the hopes that they can lipread. Some others are poring in a concentrated manner on the news of the day. In a nutshell, either the speaker should get a mic in future, or give us our handouts so we can run the experiment. Now all we have to do is get that message across.

The second session was a lot more chilled out (for most people). It was a interim casual report of where we've reached in our projects, and where are we stuck. So that actually passed by like a breeze since a lot of people had some very interesting projects.

Strategic Thinking and Decision Making

This week had quite a bit going for it in terms of a proper case discussion. We ACTUALLY read the news article/case! Well, atleast most of it, and most of us... enough to discuss! The topic was about the Challenger disaster, and what could have been the potential decision failures in the system. We go on to discuss how our biases tend to play a role in our decision making, and how we need to be aware of them. We also did a small experiment on assumptions. The prof puts up a slide, and makes us all write atleast 10 assumptions. And then he goes on to explain, how our assumptions affect our final decisions. Being the kind of people we are, we tend to abstract away and assume stuff... because apparently if we DON'T make assumptions, then we get stuck in some sort of analysis-paralysis.

So the next time you go around assuming stuff, remember... you're moving forward.

Quarter 10 - Week 7

You'd expect that as the finish line draws nearer, you'd get re-energized to run faster. I do have another quarter left (oh, the choices one makes...), and I'm looking forward to a couple of really cool courses (New Enterprise Financing and Reinvention through Entrepreneurial and Intrapreneurial Leadership) *fingers crossed for no schedule clash*, but still it's almost a feeling of relief, and the 'slow down, whats the hurry' syndrome. I want to think that I'm going to miss college once I get out, but I'd be lying. I can't wait to be done with the busy Fridays and Saturdays, the sleepless and least-efficient weeknights. Yes, I know I'll miss the refreshing gyaan (in most cases) that's being poured on us, which we then go and spout to our colleagues as if we came up with it... but hey, three years is a LONG time in an MBA programme. Let's get it done.

Business Data Mining & Decision Models

This week continues with more lab sessions... so we do Artificial Neural Networks and Market Basket Analysis. The funda of the first is that the model slowly learns how to set weights to the various inputs such that the output can be predicted. It still uses the concept of training data (the set that is used to prepare the model) and the testing data (the set on which you try to confirm if the model does indeed work). We got to see that sometimes models need to be 'tricked'. You'd oversample one particular segment of input records, to balance their weightage towards getting the final answer.

Why? Let's take an example. Let's say you want to understand attributes of the people that are likely to use an online dating service in India. So ASSUME you run an analysis on your users (OMG, you have no privacy concerns, you have no ethics etc. etc...) and ASSUME you find out that a majority of the participants like to watch TV shows about war, wrestling and cricket. ASSUME you also find that the majority would still browse through the profiles of the opposite gender, EVEN if they were already hooked up. Now you're wondering if something's off, even the so called sensible women, epitomes of loyal partners in relationships *cough*, appear to be going nuts. What you don't get immediately is this. Anyone who uses chatrooms in India knows that the ratio of men to women in such sites is *very* high. It would have been much higher, had we not had some nutty Indian guys parading as women online, for whatever psycho reasons they have. So your results are actually dominated by male attributes. Therefore, you try to oversample the women in the group, to get a more fair representation amongst the genders. Then, the results you get are more likely to be attributes of the psychographics that visit online dating sites. Get it? No? Remind me not to join the faculty at the IIMB.

The next session was about identifying what options in a set go together. For e.g. when you visit a supermarket, is the purchase of bread and butter highly correlated, or is the purchase of milk and vermicelli highly correlated (yum, payasam!). Therefore, the Market Basket Analysis tries to see if you can identify any trends in co-purchases/activity in a particular field. Fun analysis, lots of weird looking graph, but some interesting thoughts from our lab guide on what could be the possible reasons for the correlations we seemed to find. For e.g. milk and cheese. A possible reason could be that they're both being refrigerated, hence stored together...). Incredibly fun stuff.

Strategic Thinking and Decision Making

We go even further down the emotions route, we add the concept of affect this week. There were a lot of interesting principles.. stuff like how group decision making appears to be as flawed as deciding by yourself, since people tend to go with consensus more than rational debate. Or how you're likely to think more about how a particular decision benefits you rather than the company. Or why people who're advertising about 'Feed a child' always show you an EXTREMELY impoverished kid, to bias a rational decision making process. It's not a bad thing, it's just that people appeal to your emotions, instead of your logic, in some scenarios.

Or even the fact how you don't understand the scale of certain problems. For e.g. if there are two jars, a really big one with a lower percentage of green to red gems, and the other is a much smaller jar with a higher percentage of red to green gems, you're more likely to pick from the big one when asked to get more than 10 green gems in your hand (because you seem to be overawed by the magnitude). Or even when you're gambling, you're given a chance to win really big with lower probability, or win very small amount with a much higher probability, and you pick the latter because you enjoy the concept of winning more than the value of what you're winning. So the point is that the way elements AFFECT your decision making process is not visible to you, but it should hopefully be visible to others. So try to get groups to think together, and question each other, rather than letting one person dominate the discussion and then sway the group in his/her favour.

Next week's a break for Diwali, so it should be quite chilled out. I so look forward to being waken up by noisy crackers at 5 in the morning. My alarm clock chime just doesn't convince me that its worth waking up early anymore, maybe this will change that perception...

Quarter 10 - Week 6

It's been a reasonably calm week, not much out of the ordinary. The profs continue to remind us about the upcoming projects so that we don't go crazy in the last one week. They're still very optimistic. Their faith in us is troubling... or rather guilt-inducing. Whatever works, right?

Business Data Mining & Decision Models

We continue with labs this week, we've been doing stuff on K-clusters and Decision trees. The funda of the first is simple, take reams of data and let the software (Clementine in this case) work its magic on it to bundle them into clusters. The more clusters you decide to have, the more 'internally aligned' the clusters are supposed to get. So, after a little hulabaloo, you get 4-5 clusters with descriptions like (Male, under 30's, <50000$ income), (Females, over 40s, >30000$ income)... etc etc. The second session had us doing decision trees, where we feed in a bunch of inputs to the software, drag in a model and 'train' it with one part of a data set. We then just apply the trained model to the rest of the dataset and check if it still accurately identifies the type of output we're expecting (based on its understanding of the training data set). Totally takes the fun out of number crunching, but saves us a heck of a lot of time. Yay. And here I thought, the point of number crunching was to identify some human aspects of what were evidently just becoming statistics.

Strategic Thinking and Decision Making

A very creative midterm, if you ask me. The prof tries to compare a Prisoner's Dilemma with a marriage, and asks us what strategies go where. Or he asks us why Apple comes out with an iTunes store, what's the point of it? Was it strategic, or a hint from God? Some questions like that, which required us to think, assume and chew on. Finally when we start discussing the solution, heated debates about what assumptions are right, what weren't. Why one answer is as good as another... sigh, that's probably why some IIMB profs stopped getting creative with their questions. It's difficult to 'hold on' to assumptions...

We were supposed to do a case in the next session, but apparently EVERYONE forgot there was a case, and didnt even realize that the handout wasn't in the book. The perplexed prof has come totally prepared for a heated case discussion, and then realizes to his consternation that the guys in class didn't even realize there was a case (though in our defense, we had a friggin' midterm!). So already tired from the debate, and the futility of it all, the class gets cancelled and for the first time... in our time at IIMB... we were left with a free period. I should be happy! Wonder why it doesn't feel good then.

Quarter 10 - Week 5

It's the usual calm before the storm... midterms are beginning this week, project proposals are also expected to be turned in. The guys working on their final projects are frantically running around trying to give some shape to their mid-term submissions. The profs have now stopped chiding us on the fact that we're not reading the readings, and have not sarcastically taken on to the 'In the reading you SHOULD have read, the saying goes...'. Everything's normal, we're in the midst of the sea and there are no rocks to crash against, yet.

Business Data Mining & Decision Models

We're coming to the more complex data mining modes - we just began with text mining this week. The prof takes examples of how words by themselves are not as important as n-grams. These n-grams are certain combination of words that actually have a sensible meaning. For e.g. 'great' is a good word, but 'not great' implies sometihng else entirely, while 'not great aesthetics' implies yet another meaning. Hence, he goes on about a case where they had to identify trends based on feedback left behind on something. I'd have been more specific, but I'd fallen asleep thinking of n-grams.

The second session was our first practical session on Clementine... which does some pretty cool stuff with data. It's like excel, only that you can do some cool graph-y things with it. Combine data sets, filter stuff out, sort it, sample it... and finally display or extract stuff. And we got to work on a cool data set too! One of login data of students and profs.. probably to check attendance and other such stuff. Good fun...

Strategic Thinking and Decision Making

We continue down the emotion route... further refining our understanding of Game Theory. Apparently, humans are lazy, that's a shocker! They tend to make decisions based on incomplete information and rely a helluva lot on their intuition. Again, such a shocker. The funda appears to be bounded rationality. We tend to look for answers, and will continue looking till we get what we deem to be a satisfying answers, and not the optimal one. We also seem to have some sort of an opinion bias? We are more likely to think our harebrained ideas are the awesomest, while some brilliant ideas from others are downright crappy. Apparently, our intuitive part of the brain is different from the reasoning part of the brain. What makes it worse, is that we can't realize when our intuition is suggesting something downright wrong.

We're also quite risk averse! If we're getting a chance to look at good things happening, and are confronted with two options, we're more likely to take the less risky one. And if we're getting a chance to look at bad things happening, and are offered two options, we think 'what the hell' and are willing to take much riskier options. So, interestingly, the way a problem is phrased to you ends up with you taking different decisions. Good tip to have, no? It's one of the reasons we stay invested in a market that's obviously going down (because we *think* it will come back up again anyway), or why we don't invest in markets when they're going up (what if it goes down?). Interesting to know how nutty we can get!

Quarter 10 - Week 4

It's about that time of the quarter when people suddenly realize that they're supposed to settle down on a project topic, and then heated debates rise on what to study. The debates essentially come down to convenience vs. interest... 'my topic is easy', 'yes, well, mine is easier!' or the 'Man, let's study social networks!', 'Are you crazy?! Let's do a system that can rival S & P's credit ratings of countries!' to the ever glorious 'The prof doesn't understand your topic, we'll get less marks', 'The prof doesn't understand my topic, we'll get more marks!'

Business Data Mining & Decision Models

This week was supposed to be about pruning/cleaning data. It might come as a shocker to you, but people aren't collecting stats on their companies/divisions/teams so that data analysts can come and do one click with their mice with a magical flourish and charge them a bomb. Everyone wants to make everyone else work, so they deviously collect some data in one form, some related data in the other, jumble the two up and generously offer the analyst some tossed salad. If that weren't enough, they will specifically allow users to half-enter their data (in response to surveys etc.) and submit it, without offering the 'Fill everything, or you dont go further!' clause. Hence, the analyst has the mind-numbing job of having to sift through the data, find out what goes where... neatly order it, find some intellectual way of filling the missing data.... and THEN he clicks the magic button. The prof takes us through multiple examples where given a subset of data, how would one go about gleaning enough to complete the dataset. Appeared to be fun, will know for sure when we start.

Strategic Thinking and Decision Making

We had a quiz worth 20 percent of the overall count, so we justifiably had a reason not to go through the readings for the day. The prof enters the class, and is eagerly getting ready to talk of the Cuban missile crisis, and describe the various mechanics at game theory at work! And he asks 'What do you guys think??'... Pin-drop silence. Even the crickets that typically make an appearance here and answer in an unknown 'crik-crik's were silent, presumably since they were studying for the quiz as well. The prof fumes on how we've got to catch up, this is nothing, the readings are tiny compared to what we'll get later... and I'm guessing he doesn't realize that it is not really a motivating line to begin with.

In any case, we continue discussions around game theory, about how the Cuban missile crisis was an example of how the two most publicly powerful people at the time signalled their intentions to each other and how the world was at the brink of a nuclear war. Then, for the first time there was a little addition of emotion to rationality. The prof makes us play a few games, and then he questions why we don't behave rationally... and then brings in a sense of fair-play etc. to the table. Apparently, if I was offered a 100 bucks, on the condition that I share it with someone... I could come to you and say here's 10 bucks, I'll keep 90 and you're likely to show me the finger, maybe give me two rupees due to my apparently dismal state. What you don't realize is that if you don't accept the 10 bucks, you get nothing anyway. So you have nothing to lose even if I were to give you ONE buck. By choosing to walk away, you show how you bring a sense of fairness to the game. You'd rather take no money than take a pittance or unequal share. And hence, we begin our descent into the nature of the human aspects of the decision making process.

Quarter 10 - Week 3

The IIMB Vistas are coming up, and the buildup is palpable. Some of the final years have finally begun to ponder if now's a good time to take part in the fest, considering it's their last time to do so! The first years are enthusiastic, some with starry eyes... the others with a newbie enthusiasm to make their mark. The second years are busy co-ordinating, from the PGSEM side of things... to ensure that we do our bit. Pretty much like a family when you think of it... the grand-people are thinking if they've done enough to leave their mark in the sands of time, the young want to go out there and rule the world, and the stuck-in-the-middles are busy paying all the bills.

Business Data Mining & Decision Models

This course name is still such a mouthful, I have to check the previous blog posts to know what the name really is. This week we move along with more means of analysis, and we cover Artifical Neural Networks and Logistic Regression. The funda from the Neural Networks is pretty simple. Assume you have a black box, and you know that there are some inputs going in one end, and some output coming out the other. You take a sample, look at the values and arrive at what weights are to be associated with them in order to get the result. Then you take the next sample, and by applying a set of simple formulae, you find out how wrong the previous weights were and correct the weights based on the current inputs to get the output. And the cycle goes on and on, till hopefully you have such a confused set of weights that begins to predict what the output should be given a set of input values. The prof covers the funda with a case where he used such a thing, and then we marvel at how accurate something can be, and give a contemptuous smirk at all the previous methods we learnt over the last two weeks. Artificial Intelligence rules!!

The next session when we read about Logistic Regression, we find out about how this uses an exponential function to find out what the effect of the input variables can be. While explaining all of it, the prof decides to show us another case study where he applied this and how it gave all awesome results, and we smirk again at the uselessness of the other methods in this study. Funnily enough, this time the Neural Networks method was pretty bad. So, the question then arises that if all the models have an insanely unpredictable error rate, how do you identify the best one for use on incoming data, when it cant even predict past data as accurately. To which the prof sagely responds - It depends on the input format. If you use continuous variables in these two, and bins in those two, then this method is better else that method is better. That's just convenient if you ask me - Tomato / Tomaahto! A value is a value, how does it being continuous affect the model to be used is beyond me, but I didn't do the PhD or run the painstaking tests and examples, so I'll leave it to the wizened old goats and trust their statements as I've done for the last 2.5 years.

P.S. Yet another difference between the PGP and the PGSEM came forth this week - We seem to have a PGP-er attending the class alongwith us, and the chap is definitely the most inspired by this prof. How would I know? Because while his eyes are focussed on the prof like the latter's a wonder of the world, and the prof's simplest joke or statement is met with a wide grin or a vigorous nod or scribble, some of the PGSEMers are more interested in his reactions than the prof. Analytics must really be their holy grail, if they are dive this deep in a lecture. I'm guessing the prof realized he's losing our attention, because he now cracks fewer jokes. Therefore, to be mesmerized 150% or not to be, that's the difference between the young and the 'experienced'. (Bows to a rumbling applause, and a few delirious fainting acts for the earth-shattering gyaan)

Strategic Thinking and Decision Making

'Tis a game this week, and one where a number of PGSEM team look forward to tearing the guts out of one another to be the top boss! Our prof thought now'd be a good time to let us try out rules of game theory, and so asks us to be ready with a strategy to take on the other teams. And so we prepare over the week and come to class, only to realize that the rest of the class is just as nuts as we are! To put it simply, the game was five round of bidding for business and at the very first round, one of the team lands a shocker by going so deep that the rest of the people go blank for a second. From then on, we ran around trying to figure what the other crazy people might do to get the order, and we went into mutual demolition. The price of the winning tender kept going lower and lower... till finally we look at the numbers and realize exactly why this happens in the real world as well.

We then move into some theory, to understand how people can resolve the prisoner's dilemma. After hearing them, I just wonder why we need to learn strategy when obviously the answers lie in the good ol' religious tomes. Be good, care for others, don't fight, forgive people etc. are the way to go in long term relationships. It's not put as lightly as I just did... there's obviously tons of backing, because hey, it's strategy and its based on inter-entity behaviour. The prof ends with a reminder of the quiz next week, and as usual threatens us enough about reading the soporific chapters early on, if we are to do well.

Not much else on campus, other than a guest lecture on Open Innovation. The preps seem to be in full swing for the Vistas, the students seem concerned about the type of projects to choose, it's a normal week.