Business Intelligence (BI): Organizing, Categorizing and Accessing Data
- 0:06 What Is Business Intelligence?
- 1:30 Predictive Analytics
- 5:15 Online Analytical Processing
- 7:38 Lesson Summary
Business Intelligence uses specialized information systems to gain a competitive edge in the marketplace. Learn about some of the specific tools used in Business Intelligence.
What Is Business Intelligence?
Business Intelligence, or BI, uses a computerized intelligence system to collect, manage and interpret information about a business to make sound business decisions. BI monitors the operations of a company intelligently. BI employs a combination of tools, such as database management, decision support systems, data mining and statistical analysis.
How is BI different from a general decision support system? First, BI is not focused on supporting a specific type of decision but on all the operations of a company. This makes it more comprehensive. Second, BI often takes place on a continuous basis in close to real-time. This makes it possible to get an ongoing, integrated view of a company.
Third, BI has an element of intelligence, which means it tries to integrate information internal to the organization, and also information related to the market conditions and specific competitors. BI is often used not just to make well-informed decisions but specifically to gain a competitive edge in the marketplace. The specific focus of BI is often to gain more business or to improve market share.
BI often has a predictive aspect to it, referred to as predictive analytics. What will the marketplace look like in the next few weeks, months or years? What will my competitors do? What will give my company a competitive edge in the near future?
These questions are not addressed by brainstorming through pie-in-the-sky thinking but by computer algorithms analyzing data. Predictive analytics is a type of data mining that is focused on finding patterns in existing data in order to predict trends and behavior into the future. One example of predictive analytics you are already familiar with is weather forecasting. Meteorologists use long-term historic data and more recent measurements of the atmosphere to predict the temperature, wind and precipitation a few days into the future. The forecast may not always be accurate, but it is often pretty close.
A good example of predictive analytics in the business world is the insurance industry. Let's say you have just bought a new car, but before you can even drive it off the lot, you need to get insurance. You call an insurance agent to set up a new insurance policy. How will the insurance company determine your monthly premium? This will depend first of all on the actual car you drive. If a very expensive car gets into an accident, it will cost more to repair. It will also depend on how you will be using your car and where it will be parked.
Can you think of some other things the insurance company wants to know? How about your age? Definitely. Your gender? Yes. The ZIP code where you live, work and/or go to school? Those too.
So let's say you are 24. You provide all the other information, and the insurance company calculates your premium at $137 per month. You turn 25 in a few months, and you want to know if it makes any difference. Sure enough, your premium will drop to $125 per month.
So, how exactly did the insurance company calculate the $12 difference? Will you be a much better driver in a few months? Remember how insurance works. You get an insurance policy, and when you get into an accident, the insurance company pays for your repair. So, the insurance company needs to calculate what kind of risk you present to them. What is the probability you will get into an accident in the next 12 months?
This is determined by looking at the data on drivers from the past couple of years. How many drivers with a similar profile got into an accident over a 12-month period? What insurance companies have found is that younger drivers under 25 tend to drive less safe than those that are older.
Whatever the exact reasons are behind this, this is what the statistical analysis shows. What this means is that younger drivers have to pay a little more. The $12 difference is determined by the statistical model used to predict the risk of a driver getting into an accident.
Predictive analytics use historic data to predict the probability of you getting into an accident. That doesn't mean you are going to get into an accident, of course. But let's say the insurance company has 250,000 customers with an automobile insurance policy. Every year, there are going to be a number of accidents. Predictive analytics helps to determine the overall risk to the insurance company. That overall risk is used to calculate the premiums. Customers will have to pay more or less, depending on how much they contribute to the overall risk.
Online Analytical Processing
One of the other specific tools used in BI is Online Analytical Processing, or OLAP. OLAP is an approach to quickly answer questions that have multiple dimensions. You can think of a data table as having two dimensions: rows and columns. For example, each row in the table could be a product made by a business, and each column could be how many units of each product were sold every month. You can answer questions about this table using a database query.
OLAP extends this idea to more than two dimensions. For example, for each product you also want to know in which facility it was made (3rd dimension), which customer it was shipped to (4th dimension) and how many units have some type of problem as reported by the customer (5th dimension).This information could be stored in a relational database, but for a large number of dimensions this can get complicated, and the database will become slow. OLAP organizes this multi-dimensional data, so that it can be analyzed quickly.
For example, once you have the data organized in the various dimensions, you can ask questions like this: Show me all the facilities in Europe that made this particular product in the month of June that were shipped to customers in Asia that were reported to have a problem. Now for those same facilities, show me all the other products they made in the same month and that were shipped to North America, and then give me a mailing list of all those customers in California.
OLAP is designed to process this type of analysis very quickly. You can see how OLAP can be very useful to analyze complex business operations. OLAP and data mining are both used in business intelligence, but they represent different approaches.
In data mining, analytical tools are used to uncover relationships in the data. You are basically saying, 'Here is the data, show me what the interesting patterns are.' In OLAP, you start with some very specific questions, and you are drilling down into the data to find the answers. Data mining is bottom-up, discovery-driven. OLAP is top-down, query-driven.
Business Intelligence, or BI, uses specialized information systems to gain a competitive edge in the marketplace. BI employs a combination of tools, such as database management, decision support systems, data mining and statistical analysis. Two specific tools, which are somewhat unique to BI, are predictive analytics and online analytical processing.
Chapters in Business 104: Information Systems and Computer Applications
People are saying…
"This just saved me about $2,000 and 1 year of my life." — Student
"I learned in 20 minutes what it took 3 months to learn in class." — Student