Monday

Life of DW-BI

What is Business Intelligence?

  

 

Simply put, Business Intelligence is a set of thought, money and actions put towards either saving money or getting more money for the business. How do you do this? Obliviously by analyzing some data with some tools.

For example, consider yourself as a store manager of Nike who wants to put up a shop in a new city. So, where should you put up your shop? In downtown area? Or near a college? or near some sport equipment shops? If you are a good store manager who wants to make more money you would consider analyzing some data about demographics, competitors, property rates to come to a magical combination which will cost the least but will drive the sales through the roof.

The process of transforming this input data into actionable insights will help you to make strategic decisions (like finding a sweet spot for your shop). And this my friends is termed as Business Intelligence..!! Below mentioned info-graphics will help you visualize this process.

Typical Life Cycle of Business Intelligence Application

It all begins with data! Whether it be structured data like relational tables, semi-structured XML files or unstructured data like facebook posts, it usually goes through ETL. Here data is extracted from its source and placed into staging area where its cleansed, transformed and loaded into Data-warehouse. The next phase is reporting phase which gets its input from data-marts. This information is then visualized using reports or dashboards and is also made available across mobile platform.

Now, lets understand some more key concepts by continuing our Nike Store example.

Data-Warehouse


Data-Warehouse is like a neatly arranged and organized closet which gets you a quick access to your day to day clothes, keeps track of your dirty laundry and tells you that its time to buy some new clothes. In formal words- A data warehouse is a subject-oriented, integrated, time-variant and non-volatile collection of data in support of management's decision making process.
  • Subject-Oriented is similar to keeping your regular, party and sports wear in different compartments.
  • Integrated is like having multiple sources for your clothes- newly bought, borrowed from friend or passed on to you by your older sibling.
  • Time-Variant means keeping some very old clothes- your first football jersey or your first Halloween costume.
  • Non-Volatile can be termed as once you get clothes you never alter them. If you are growing or getting fat- you just get bigger sized clothes
All these properties of an ideal closet (or data warehouse) will help you (the manager) take more informed decisions.

OLTP v/s OLAP

Online Transaction Processing a.k.a OLTP

Consider you working at the point of sales (typically where you do the billing) of your new Nike store where:
  • You have to process hundreds of billing transactions in a day which get stored in some operational database. At the end of the day you simply query daily sales against this database to verify the cash in hand.
  • The system should be fast and order info must be inserted, updated and sometimes deleted (in case of incorrect order entry) within seconds.
  •  You just worry about handling the fundamental business process - take cash in and sell goods to customer. This is a Online Transactional Processing (OLTP) system.

Online Analytical Processing a.k.a OLAP

Now some years have passed by and your hard-work has paid off.. and now you are the manager of the store where:
  • You do run some complex queries to analyze the state of your business. How are your monthly sales compared to last year? How much discount can you afford to attract more customers for the coming Christmas sale?
  • Obviously these queries will take some time time to run because you are consolidating data for few months (not just doing some insert, updates and deletes)
  • Now, your aim is to take some informed strategic decisions which can enable you to increase your revenue further. This is a Online Analytical Processing (OLAP) system.

ETL - Extract..Transform..Load

Extract, Transform and Load.. This is how the data warehouses are built.!!
  • Incoming goods for your Nike store are first extracted (or received from one or many locations of the  company factory). 
  • These goods are checked for any obvious defects (like a T-shirt having only one sleeve). Goods in good condition are packaged and labelled with the sell price and discount information as per the local currency (transformation). 
  • And finally these transformed goods are loaded into appropriate sections of the store (like Men's, Women's, Kid's etc) sometimes these sections can be termed as Data-Marts in data warehousing terminology.

Reporting & Visualization


 

So how do you decide if your store is doing well?
  • First, you decide on some KPIs a.k.a Key Performance Indicators. So, in case of a Nike store these KPIs can be number of customers, number of orders, total sales, total profit etc.
  • Once you have the KPIs you usually would like to see them in nicely formatted reports. These reports can be Daily Sales report or Quarterly revenue report.
  • If you own several stores in the west coast region, you are more likely to see overall performance of these stores with respect to time and location. That's where Dashboards come in.
  • And if you want to compare you store sales with your last year's sales.. or you want to compare actual sales v/s the forecasted sales then you probably want to have a scorecard.

Lets hope this new role as a Nike store manager helped you to refresh the understanding of common BI terms. Stay tuned for more interesting posts on Data Analytics and Big Data! Feel free to express yourself in the comments section. Happy Learning!

~Slice of BI