Business+Intelligence+Tools

**Business Intelligence Tools**
Sunday, November 13, 2011 <>

**Topic overview:** Once you have succeeded in gathering together large quantities of high quality data in your organization's data marts and/or data warehouse, you need computational tools to help you extract valuable information and knowledge from them. //Business Analytics// describes a broad selection of computational tools and techniques for analyzing and making sense of these large data collections. In class today we will introduce some basic principles for the effective selection and deployment of business analytics tools, as well as take a quick look at a variety of specific tools to show how they can be applied to answer questions about your business, or even to help you pose questions that you had not yet thought to ask.

By the end of today's class you should be able to:
 * Understand the role of a data warehouse in powering business analytics tools
 * Identify and describe a variety of different business analytics tools including tools for: reporting, data mining, data visualization, performance dashboards, and web/text mining.
 * Explain the concept of a Key Performance Indicator (KPI), and be able to specify precise KPI's that would be useful in real-life business situations

**Preparation for class:** In preparation for today's class, you should read the following two articles from recent issues of the McKinsey Quarterly journal:

These articles are available at the links below, (site registration required), or as PDF files on digital reserve on Blackboard, in the 'Course Materials' section.
 * Brad Brown, Michael Chui, and James Manyika, Are You Ready For The Era of Big Data?, McKinsey Quarterly, October 2011.
 * https://www.mckinseyquarterly.com/Strategy/Strategy_in_Practice/Are_you_ready_for_the_era_of_big_data_2864
 * Competing Through Data, Three Experts Offer Their Game Plans, McKinsey Quarterly, October 2011.
 * https://www.mckinseyquarterly.com/Marketing/Digital_Marketing/Competing_through_data_Three_experts_offer_their_game_plans_2868

In addition to these articles, you will likely find it useful to spend some time experimenting with the sample Business Performance Dashboards at SAP's Business Objects / XCelsius website:

http://www.sap.com/solutions/sap-crystal-solutions/dashboards-visualization/demos/software-demos.epx

There are about 14 sample dashboards on this page. The examples that I found most interesting include:
 * Daily Executive Report
 * Analysis By Regions
 * Sales Visual Model
 * Loans By Branch
 * Competitive Comparison Dashboard

**Reference materials:**
Slides:

In-class exercise for KPI's:

**Pre-class exercise:** Select a company with whose products or services you are familiar, either as an employee of the company or a customer of that company. Describe a specific business challenge that company faces and suggest one concrete, specific way that the company might use data mining and/or predictive analytics technologies to address that challenge.

Try to come up with examples that other students (and the instructor) have not already posted. Taking what is basically the same scenario as a previous post but changing the company involved doesn't really count as coming up with a different example...

[Bob Monroe]

=
Toyota sells its vehicles with warranties that require the company to repair vehicles that have defects within the warranty period (typically three years) at the company's expense. It is, therefore, obviously to the company's advantage to prevent the vehicles from having expensive breakdowns. One way that Toyota might use data mining and predictive analytics is to help Toyota mechanics and service centers diagnose and prevent expensive repairs before they occur. To do so, they could run data mining tools on historical service data to see what symptoms and repairs were most commonly performed on vehicles that later came in with very expensive problems to repair such as broken transmissions, overheated engines, loss of engine oil, etc. The data mining tools could discover common patterns of repair work that preceded and possibly predict later major repairs. If there are simple precautions or checks that the mechanics can do early to prevent these major repairs occurring at a later time, then the predictive analytics tools could recommend these inexpensive precautions and preventive actions to the mechanics and the customers when they observe that the pre-requisite conditions have occurred that predict a later major breakdown. =====

By reducing the number of vehicle breakdowns, Toyota will potentially reduce their warranty repair costs and also improve customer satisfaction with their products and services.

[M Hammad Abbasi]

As an intern at KPMG, one of the problems the HR branch in KPMG faced was in recruiting. There were too many applications coming in, and there were many universities in Pakistan to choose from. KPMG could not make a presence at all of them and so they had to choose which university to focus their recruiting effort on. In this case, data-mining can be used to identify characteristics of their best employees such as quantitative skill, knowledge of Excel, years in high school, which university attended etc. Once they use data-mining to find significant characteristics, they can then use such techniques to determine which universities and which regions best fit these characteristics and focus their recruiting efforts there.

[Walied El Hag Ali]:

One of the problems that could be rectified using data mining tools or predictive analytics technologies at Lulu Hypermarket is solving their demand issues and making expansion decisions. I believe that this problem occurs due to bad research. This is because at some areas we find that Lulu is packed with people at all times and at other branches it is relatively empty. In this case, data mining and predictive analytics could be used on information about Lulu's current demand and also on their historical records. This will help them identify the branches in which they have high and low demands and then expansion decisions could be made accordingly to solve the issue.

[Nahan Arif]:

I did a summer internship at Qatar Airways and one of the most critical drawbacks of Doha International Airport is its small size. Early morning flights and late night flights are the peak time hours everyday where many customers have to queue up for long hours at check-in and at the gates. To avoid negative publicity of the long waiting hours, Qatar Airways can manage demand by taking advantage of data mining and predictive analysis. They can do this through doing analysis of how many customers on average fly to what destinations, and which of these destinations would require more than one check-in gates to avoid long queues and missed flight due to the long waiting hours. Data mining tools will be able to discover the common patterns among the peak times and the number of passengers.

[Maryam Al-Thani] One problem that many Abaya shops in Qatar, and the Gulf region have is the lack of a customer database. Topaz is one of these companies, when you call them to follow up on your order they simply have one order number associated with your order and do not have any historical data. Since many abayas are customized to customer's needs in terms of size and style this creates a problem for the customer as they have to provide their size and preferences every time. Once companies such as Topaz have created a thorough customer database. They can use data mining techniques to check historical records of customers and their preferences and can notify them when they have abayas that they have new designs that they can create for specific customers.

[Khadeejah Al-Husseiny ] In high school our lunch time break was a mess. All grades from kindergarten uptill 12th grade had lunch hour at the same time. There was only one cafeteria that served food, so people had to compete to get inline first, or wait pateiently but risk that by the time they got their food it will be time for class. The worst part was that younger kids wouldn't even eat because they were scared to get crushed by older kids. I believe that that one way the school can use data mining techniques is by organizing their data such that they can figure out when teachers teach what class and how many classes they teach. They can then reorganize the class schedules in order to have lunch time at different time so that there is less traffic in the cafeteria and more people will get the chance to order, and the cafeteria gets more students (customers) to serve

[Mohammed Kamal] Apple produces the Ipod in 3 different memory sizes, the 8GB, 16GB and 64GB. Apple can use data mining techniques to help them in the manufacturing stages of the Ipod. Analyzing large data from the apple stores all over the world, Apple can figure out which memory size is selling more than the others and thus produc

[Al-Jawhra Al-Mana] Qatar Airways flies to countries all over the world, all year along. But at some seasons or certain periods of the year, some of the flights are fully booked to specific destinations, and customers need to book 3 or 4 months in advance for the flight to get a ticket especially during summer vacation. Qatar Airways can use predictive analytics technologies to assign the right number of flights and the perfect size of airplane to each country to maximize the number of tickets bought and increase customer satisfaction. In order to do so they need to know the countries that are most visited and during what season it's flights are fully booked and the other countries that are rarely full, so that Qatar Airways can make the right decision for choosing the number of flights to each country.

[Maryam Al-Subaie] As a customer of FAO Schwarz, I was added to their mailing list after being asked for my personal information (including email and home country address) at the cash register in the store. The company used this as a channel to communicate all about their new offers and promotions such as ‘Shop For a Cause’, ‘Back to School,’ and ‘Halloween.’ These were clearly irrelevant to me because I don’t live in the States so I wouldn’t be able to visit the store regularly to see their merchandise etc. Secondly, their promotions were directed at parents and not buyers who were purchasing gifts for other children. The frequency of their emails, especially around the Christmas season can become annoying that customers will disregard them - I directed their emails to my spam and ultimately unsubscribed from the list. They could have used data mining tools to analyze geographical locations of customers first; it would have helped them aim specific offers at specific buyers. For example, they could have emailed me information or updates about their online store/shipping details etc which is something I might use later in the future.

[Hind Al Khulaifi]

When I interned for HSBC, I was responsible for making spreadsheets that pinpointed certain trends in the performance and accounts of the bank. Because of data mining, I was able to have all the necessary data I needed for customers of the bank and their details. By filtering the accounts that were deliquent more than one time, I was able to notice crucial trends. One such trend showed that more than 70% of a certain nationality were deliquent more than four times, therefore such important piece of information is crucial for the bank. The manager had told me that the previous year they established new restrictions to giving loans to a certain nationality becauase most of the customers never ended up paying their fees. That way the bank can gaurentee that no major losses are incured in the future because of this.

[Layal Al-Alami] Like many CMU students, I believe that Bateel has a supply/demand mismatch, which costs them customers, and sales. Bat eel can solve this problem by keeping track of what items people order more, and what items are usually left not bought. This way, they can get more of what people really want. This way Bateel can sell more items to student.

[Abdullrahman Al-Muftah] It is without doubt that hotels in london have thrived over the past few years. One of these hotels is the Mandarin Oriental situated in Central London. After doing an internship with KPMG in London, I had the opportunity to meet up with Mandarin Oriental's top managers whose main task was to be in charge of customers who had come from the Middle East. I had learned from this meeting, that over 30% of Mandarin's customers came from the Middle East, and that this number was increasing every year. Although the Manager was happy with this number, he wan unclear on why large number of people from the Middle East would "flood" the hotel at one point. Mandarin should analyze their customer's behavior, analyze patterns on when people from the Middle East are likely to come, and why. Mandarin also offers its customers several promotions, which includes thanksgiving and christmas promotions to its clients. Although these are important holidays and events that are celebrated, the Middle East also celebrates Eid Al-fitr, Ramadan, and Eid Al-Adha. Mandarin could create promotions specifically for these time periods, and then carefully observe the outcome of these decisions. Finding this information out, can increase the number of customers from the Middle East, and in turn the hotel's profitability.

[Ayah Abujarbou] when shopping people usually find it difficult to find the right item that matches their need. Thus, clothing stores for example Zara should see what items do customers buy and if the same customer happened to buy multiple pieces of the same item but different colors that shows that they liked it and in the future they might even come back and get a new one of the same item. thereby, stores might look at their database and consider whether they should include that item in their basic collection that is always available.

[Hussain Hejji]

Last summer I was an in tern at Qatar Shell. Part of my job was to get the details of the people clearing the parked invoices. Because of data minning, I was able to easily track the people clearing the invoices and how many they have cleared over time. I think this was an excellent way of using data minning for tracking.

[Fatima Al-Khayat] Adidas one of the biggest sportswear manufacture in the world lately have decided to produces kits of famous club that they sponsor like Real Madrid Club without the players name on the kit (shirt). Their claim was that kit with player name that are not famous or popular doesn’t get so much sales. In this situation, Adidas has recognized the problem, but took the wrong action. Since they can identify the popular players, they could produce more kit with the popular player name to increase their profit and satisfy their customers.

[Noor Al-Mohannadi] Aramex provides a transportation service that enables regional and global trade. Recently Aramex have misplaced two of my packages, one was in the warehouse and the other one they gave it to the wrong mailbox. To increase and maintain their efficiency Aramex should find the main reasons behind such incidents and later try to find solutions to this problem. This will help Aramex avoid losing packages and paying refunds to their clients.

[Amal Osman] ==== Carsales website, they are not utilizing and using their resources effectively, the information on their website is not organized well enough the search engine is not that beneficial since it does not sort out to help that you need. Data mining can help the company find the patterns of the demand and patterns in the customer preferences and and their peaks of sales. So for-example they can figure out that at a certain time people buy a lot of family cars so when that time comes, there could be a list of family cars that would pop-up on the website, and that would increase their sales. ====

[Mohammed Al-Rawahi]: In order to reduce traffic accidents, Ministry of Interior can record the place, time, cause and frequency of traffic accidents throughout the country. Then when having this information, actions can be taken to reduce the frequency of accidents. For example if the data show a high number of accidents in a road which is a result of insufficient lighting in the road, it may consider to increase lights in that road. Data mining can be very helpful in order to reduce accidents.

[Fatima Abdulla]: I did an internship in Doha Bank, and I was responsible of clearing and also tracking the checks that comes to the branch. Because of predictive analysis that they conducted, they had information of the most famous banks that people deals with them in other countries, which make it easier to people to know to which bank they can connect with in other countries for the transformation of money from this bank.

[Ahmad Al-Sarraf]

Restaurants, in general, should use data mining to forecast the demand of each dish they have to make each day. Few months ago, I went to an Indian restaurant, Asiana and ordered my favorite meal, lamb biryani. However, the waiter told me they don’t have lamb biryani and the only biryani dish they had, at that time, was shrimp biryani. Even though it was 7 PM on Tuesday and the restaurant was kind of empty, they did not do a great job in forecasting the dishes for that day. That is why it is important to use data mining to avoid these kind of unexpected demands.

[Hamsa M Al-Massri] Curves health clubs and fitness centers come across a variety of cases everyday in their branches all around the world; for example, if person x is a female that has a body fat percentage of 10, weighs in the range of 45-55 KGs, etc.. then they should follow a specific diet and do specific exercises. I believe it is very important to use data mining in their cases to be prepared for all the daily cases they come across; to provide the best results for their customers, as well as, becoming more efficient and effective in what they do.

[Sara Al-Mannai] Jarir bookstore can use data mining to track customers buying behaviors, or what people here in general prefer to read to other books that they are not very popular to their customers. If you go to Jarir, there is a wide range of books that are not always sellable or very popular. So instead of the store having to shelf all of the books they can find, they can use data mining to see what customers prefer whether genres of books or authors they can bring more books from to further satisfy the customers and make it easier for them to get the right books instead of having the wrong books that almost no one buys and end up on the sale racks.

[Anas Ali Chaudry]: I was working in Bank Alfalah in Pakistan during the summers. There was a problem with the recruitment. The office remained understaffed in the customer relation department and when people took vacations it became more hectic for others. For this they can use data mining to predict number of customers coming in at different period

[Nasser Al Khori]: This year during Hajj, the Kingdom of Saudi Arabia have introduced a new means of transportation which was the train system to move over 4 million people from one place to another. The problem with the train system was that people had to wait for them and that the wait was something around an hour or even more. Moreover, another problem existed was that I observed some trains arriving at the station mostly empty while others totally packed. A solution to this problem is that there should be a data mining system to forecast the demand for trains, during certain hours of the day. Data mining can be extremely helpful to eliminate traffic and people having to wait.

[Najla Al-Madhadi]: Almeera Consumer Goods Company can use data mining to analyze local buying patterns. Due to the fact that Almeera has many branches throughout Qatar, Almeera can use data mining to know what products mostly demanded in each branch. They can also use data mining to track the products demanded in specific seasons like the Eids and Ramadan to provide them to their customers.

[Haya Al-Mannai] during last summer I worked at ICT Qatar, during this time I observed one thing, that some employees at the HR took vacations while there weren't enough employees to replace them during this period of time, which made the work overwhelming. So, data mining will be very helpful in this case so that they can predict the amount of work they will have during different period of times in order to know when can they give their employees vacations and when they cannot.

[Nijat Ibrahimov]

Some Petrol Stations have a demand-supply mismatch when they have excessive customers at a certain times when it leads to frustration of customers and employees. Obviously, petrol stations might want to avoid such situations and not loose their valuable customers. One way of doing it, management of petrol station might use data mining and predictive analysis is to avoid such demand-supply mismatch. Obtaining data and recording it in different times can help petrol station to know when they need to have more employees on duty and how many customers leave petrol station not served. This is vital piece of information might explain management how to deal with this problem. When they need to allocate more employees or even expand petrol station on order to increase flow rate of cars.

Solving demand-supply mismatch problem for petrol stations will make profits of company at will lead to the customer satisfaction. [Firas Bata]: Carrefour is the world's second-largest retailer. They sell a large variety of products from food and household necessities, to electronics and mobile phones. Having said that, Carrefour could utilize some data mining tools to analyze and understand their consumers' buying habits and patterns. This would help Carrefour's management make better decisions in terms of product placement as well as predict sales of certain products that some consumers consider to be complimentary products. [Laura Jaber] In Lebanon, we have a medium size shawerma place that gets a lot of traffic. I can't remember the name, but we go there a lot, so i have noticed that around 5 pm, they always run out certain dishes like meat shawermas and roasted chicken. After 5 pm their traffic usually thins out because of this fact, leading me to assume that people have noticed this fact and don't bother to go there after 5. This causes them to lose customers and so i would suggest they make a system to keep track of the selling patterns and from the patterns find a solution to keeping stock of the dishes that run out due to high demand. [Mohammed al ahmadani]: The traffic department in Qatar should use data mining in order to fully understand how accidents here in Qatar occur and why. This mechanism can help them and the many driving schools here in Qatar emphasize these causes of accidents when it comes to teach new drivers how to drive and doing so safely. If it is caused because of the way a set of roads are made for example, then they could use the information to make sure to fix whatever is making the road unsafe for drivers. [ Dalia Hassan ]: In Qatar there are many beauty centers everywhere, but the problem that there are a lot of centers which are located next to each others. So we can find three or four beauty centers at the same street. In this case, beauty centers should use data mining to choose the right location. Moreover, they should use the data mining to have the right services that customers need instead of having many services with many employees. [Mohammed Hadi Takiddin]: At most of the petrol stations there are Car Wash section, but some of them don't have this section which needed daily. I think the petrol station must use data mining to know the need of the current location. They should know which location doesn't have petrol station and which has Car Wash section and which hasn't.

[Orkhan Rustamzade] : Ebay faces a lot of problems with its customers when they pay for the items but don’t get the item either as they wanted or they don’t get it all. Ebay should develop certain frame for new sellers who have to provide more information about their location, credit card number and id. This might violate someone’s privacy however it will drastically reduce the number of frauds and thefts. [Dua'a Althabatah]: I suggest that "faces" shop for makeup and cosmetics would use data mining and predictive analytics technologies to figure out a pattern of each customer's purchase. What I mean is to have a pattern for all the makeup brands that a specific customer usually ask for and buy. This will help them to serve their customers better and satisfy them more.