Despicable Me!

Blogging, yes, that is how desperate the tiding of my life have made me. I remember starting this blog to provide a vent to my effervescent personality and here I am, a circle completed. I am closing down on the 25th year of my existence and the concept of life still eludes me. I know I don't want to sound all philosophical and shit but it is true I am lost at sea. My old friends would remember my self composed saying from class 11th- "Life is a progressive realization of your chutiyaaps", I dont know if I was stuck my brilliance at the moment of it's composition or is it truly is the sad reality of existence. Maybe I grew too fast, matured to early to suckle the trivial joys of life that to me seemed trifle. My critique nature has become a poison. I feel the purpose missing, the motive fading and the journey tiring. There is a parasitic conglomeration of multiple personalities within me each trying to drift me in a new direction but like all symmetric vector forces they cancel each other and I stand unmoved untouched. The yonder days are too deeply engraved, I miss them. Mom always used to yell at me for having too many friends and too involved with them, but guess what here 10,000 miles away in the so called Land of Dreams with respectable money and a prospective career, them is what I yearn for. Enough of self loathing, I need a purpose, a rejuvenating force to propel me away from this carcinogenic surrounding called my mind.

MIS 587 comes to an end

MIS 587 comes to an end.... It was a busy semester starting with GOMC to various tools and softwares and then ending the semester on a high with Social Media project. Over all the entire course was very interesting, engaging and learning. We learnt the various strategies employed in the industry right from collecting data cleaning it, finding the business value hidden in that information. The course has prepared us not only for the current industry but even for the future by making us adept with online marketing and web analytics. Thanks to the various projects we did in the course I can talk a lot more about BI in the interviews than before. It has instilled a sense of confidence in me and has made me competent in facing any position and any profile. The final aim of someone wanting to gain experience in the world of BI is to become a data scientist. We have now the theoretical knowledge that we need to leverage with the industrial experience and prove our mettle. Thank you Professor and my awesome team mates.

Infographic Resume

HW-6 was one of the more fun projects we have dine in MIS 587. We have been asked to make an inforgraphic resume, basically use pie charts, bar graphs and timeline to represent how you have grown professionally over time and make this resume as an application for a job applications. So what I did was use a pie chart to represent my level of expertise and my past work experience. Use a network model to list out all the tools and programming languages I know of. Also make a tree structure to details the work I have done in the past. This was a fun project as it allowed us to use out imagination and still use concepts of BI such as data representation skills, data visualization techniques to showcase ourselves in the best way possible. Having a colorful, dynamic pictorial resume looks very attractive and maybe help us get a job that we could not have gotten otherwise. Link: https://docs.google.com/file/d/0B-Nt4sKnLiuGMUpDLWp1T0tVQ3c/edit?pli=1#

Ad Hoc OBIEE

As I went through the OBIEE tutorial there were various things I wanted to implement which I couldn't figure out from the processes explained in the tutorial so I did some self-research and found something interesting using which I could do Ad Hoc reporting!!!
There is an option of setting up a prompt in the dashboard. How the process work is that you predetermine the column or columns over which you expect the user of the dashboard would want to use his discretion. You can select the possible values of those columns he can choose and then after selecting multiple columns and multiple values in each column you can add this prompt to your dashboard.
The benefit comes when the user wants to select some or all columns to view the data of or plot in the graph. This benefit allows the user to compare a set of value against the other set of value. And use the selected view to extract business information.
The universal benefit of this approach will help a single dashboard answer company wide queries. The same dashboard can be sent to all managers whether in the same department or different, they can choose to see the various products of their department and also compare their department's performance against some other department.
The one stop ad hoc reports are the real benefit the OBIEE provides, ANd I am glad I used them to the best in my homework 5

OBIEE

The recent project was really helpful in understanding how BI works in a current enterprise whether be it healthcare or retail or forecasting based on historic data.
The tool OBIEE is a really helpful OLAP tool where you can create almost any report you want to create. The dimensional model of the data eases the complexity as the facts and dimensional tables really help one understand the details he/she is looking for and pull the necessary details to present the output as desired.
Whats improves the tool is the ease with which a person can plot various graphs and not only plot them but change the various components to get the exact data representation they are looking for. With the wide variety of details provided be it a pie chart, line graph or anything else, along with being able to swap the columns as X or Y axis is also a very convenient way by which one can present his manager easy decision making capability.
The best part is the mobility component of the whole process. You don't necessarily have to be tied down to a system and the connected server to procure the data you desire.
This is the future of BI reporting on the go!! And with the Oracle's latest service which allows users to pull reports from their iPads and other tablets makes the market more futuristic.
There is still more to come like Real time BI. Well we are still chasing our dreams and real time BI is not the end of it, as every distance you cover shows the path that lies ahead.

Data Warehousing, Star Schema, facts, Dimensions

Data Warehousing:
Shifting data from transactional system to a data warehouse or a data mart allows you to secure your data and also gives you the functionality to retrieve trends and other business values from the historic data.
Business Intelligence would not be possible without data warehouses. But even data warehouses need to be tweaked to make them best suited for BI. The process is converting normalized tables to de-normalized form so that a dimensional model can be implemented in place of the typical ER model.
Dimensional Model allows for easier data retrieval and any user even with minimal technical knowledge can understand how data is saved using star schema.
Star Schema is composed of facts and dimensions. Fact is the numerical data field which is usually aggregated over the related fields. E.g. Sales
Dimensions are the related fields that give additional information about the fact field. E.g. ProductID, StoreID etc.

Indexing allows faster query fetching but slows the system by slowing update, delete append!!!

Long Tail and e-Commerce

When I first saw the Long tail graph the first thought that came to my mind is this is 1st quadrant 1/x. and the unique property of 1/x is that it tends to 0 at infinity implies it always has a finite value for even the largest value of x. Now moving from graphical to the business implication:

Long tail property allows e-commerce business which have low inventory costs and offer low search costs to provide "unpopular" products. By just observing you can see that one can provide equal or more sales by offering low demand products than by selling the popular products.

E-commerce provider low search, storage costs which has resulted in exemption of Pareto's 80:20 rule- 80 % of your sales will come from 20% of your products. Now companies like Amazon and Netflix can house the products economically at a centralized warehouse and offer a wide range of products. This change has affected people's preferences as earlier they could not find the product of their choice which was outdated.

Final verdict: The more extensive the data warehouse will become the thicker the tail will be.

Networks and BI

Networks have become a very integral part of our lives. and technology is making full use of networks to make our lives easy. Facebook makes use of networks and lets you designate some friends as "Close friends", you are send special notifications when these friends update their status or post a new album.

LinkedIn and Twitter use networks to analyze your level of association with others and based on that you get recommendations to follow similar trends or add new friends.
But Amazon's use of networks combined with Business Intelligence is something that others have not being able to emulate.

The recommendation system system of Amazon suggests you new products based on someone else's shopping experience. and to everyone's surprise the recommendations are quite helpful.

As time progresses we will see closer association of network study and business intelligence.

Various similar terms

While browsing through the web for proper understanding of the semantics of various online marketing terms this is what I learnt:

Cost per milli=Cost per 1000= Cost per impression:
These three terms all mean the same. This model of online marketing generates revenue in the ways similar to the traditional marketing- amount of exposure. The firm is charged on the basis of number of people to whom the content was displayed. Mind it DISPLAYED.

Cost per click: Is the next step in online marketing evolution. Here the firm is charged either a fixed or a varying amount for every click on the ad being displayed. This is a better quantifying metric to estimate the success of ones marketing campaign as you can actually see how well you ad campaign is doing and accordingly change it.

Cost per action: This is the final step of evolution in the online marketing strategy. Here is the firm is charged on the basis of execution of a pre-defined action like purchase of a product, filling of a form etc. This ad campaign though best for the advertising firm is not the best from the website's point of view. because majority of the people still buy things offline (for reasons click here- http://www.theatlanticwire.com/technology/2011/12/we-still-shop-mall-bunch-reasons/45738/ ). People come online survey the product of their choice and still buy stuff from the store/mall.

Lastly we have another form of marketing campaign but this one is different in its essence. Pay per sale followed by Groupon and similar businesses needs equal dedicated efforts from both sides.

online v/s offline marketing

The times have changed. The proverb- First impression is the last impression, cannot be truer. The attention span of an average person is so small that as a product either you make it big or you find your place in the basement. In such volatile times marketing has reached a completely new dimension, with a plethora of options the marketing team had to find the right audience and target them in the only best way possible.

The technology is supporting these advancements. Benefits of online marketing
1. targeting the right person: based on the users shopping/ browsing history you can find the right person to market to rather than spend millions on trying to woo everyone.

2. dynamic: you can change/ tweak you marketing campaign very readily from the comfort of your home/ office rather than changing all the billboards all over the country or coming up with a new print/media ad, shooing for a commercial etc.

3. ubiquitous: with the huge number of mobile apps and internet devices all around us online marketing is able to penetrate every minute of our lives rather than just one glance from the morning newspaper.

Google has become the leader for these and many more reasons.

A/B testing

Testing for web analytics: testing entails one of the major costly operations for any organization. And when the issue is of website designing where user input and perspective is of paramount importance A/B testing is the best route out. A business can assess its competence at converting its visitors to customers by defining certain metrics or KPIs and then make "minor" changes to the website to see which option helps them achieve their KPI better.

The problematic issue with A/B testing is that the changes made to the website should not be drastic and frequent, and that KPIs should be well defined and quantifiable. Implementing A/B testing is easy but understanding what you want to achieve out of it and planning around it is mind-boggling.

In today's fast paced time the human mind is accustomed to living in their comfort zone, changing site drastically will offset this convenience and the A/B testing can result in a business loss.

BI for Currency Exchange Predictor Model

One of the most economically lucrative idea: Currency exchange predictor. What if you knew when the dollar price will go up and by how much, when will it fall. You can earn and pretty much make a living out of this knowledge. And you know what is behind all of this BI !!!
Using the basic knowledge of currency exchange, some awareness of world affairs and guts to invest money is all you need. I had done a very detailed report back in India but sadly it seems I have lost it. Here I have collected the 10 years of U.S. $ data from::
http://www.bankofcanada.ca/rates/exchange/10-year-lookup/

and made something very trivial but on similar lines.
As we see a periodicity in the trend of how US dollar changed over time. Anyone can capitalize on this information and earn bountiful profits.

Industrial Use: I will take my own example here. My company was based in India and our client was GUESS, US. My company used to get money in dollars and it had to pay us in Indian Rupees. This knowledge of when to get the maximum return from our $ earnings made my Vice President put this project infront of us.

The difference

My experience with the BI class till now has been great. I got to learn a lot about data collection from online traffic and the various KPIs and metrics that differentiate a better website from a worse.

But in the last class I felt there was something lacking in the presentations. Everyone
has great data to backup their conclusions but the conclusions were more data-centric rather than business-centric.

To me the picture of business intelligence is composed of data, the information we collect from the data and the reporting where we suggest ways to improve the business. The 3rd component is the most important part of that picture, because even if you have the data and you extract the correct information from it unless you can relate it to the business everything is in vain. Its like watching a suspense movie but without the ending.

The "So what?" question is something I and my classmates need to keep in mind in our future analysis.

I and BI

As an Electrical and Electronics Engineer from BITS, Pilani when I got an opportunity to work for a start-up, Apex Decisions, Bangalore India that works in BI I was confused, because before that day I did not even know what Business Intelligence is.
To my surprise I saw the diverse application of BI in all walks of life. My project area was Retail market with US based clients as GUESS and BigM. I was surprised to see all the sales and promotions run by various retail brands like GUESS, Nike, GAP etc have proper quantitative decisions behind every move. To my surprise I was told that our pilot run with GUESS had increased their yearly revenue by about $20 Million- all because of BI. It was unbelievable for me but as I started working with them and implementing their algorithm and using their tools I could feel how BI can augment a retail business.
Extracting seasonality pattern information from traffic pattern of sales round the year, considering special events like Easter, 4th of July, Thanksgiving and Christmas, and using the tools in the best way possible I could forecast sales patterns for my clients for the coming year.
The biggest surprise was to when my analysis was actually forwarded to GUESS and implemented. The accuracy of my forecast coincided with less that 5% error with their actual yearly sale. This was the moment that ushered me towards BI and couple of months later I find myself in the BI class. Hope my journey takes me to more interesting places. Later.