Monday, April 17, 2017

Chapter 9: Business Intelligence Systems



Summary: This chapter focuses on business intelligence (BI) systems: information systems that can produce patterns, relationships, and other information from organizational structured and unstructured social data as well as from external, purchased data. Another source of knowledge is employees themselves; a large amount of collective knowledge exists in the employees. BI is the key technology supporting such marketing technology.
Q1: How do organizations use business intelligence (BI) systems?
Five standard IS components are present in BI systems: hardware, software, data, procedures, and people. The boundaries of BI systems are blurry. BI is used for collaborative tasks: Project Management, Problem Solving, Deciding, and Informing. Falcon Security could use BI to determine whether it could save costs by rerouting its drone flights. What Are Typical Uses for BI?: (invoves classification/prediction) Identifying changes in purchasing patterns. Ex. Important life events change what customers buy. Entertainment Ex: Netflix has data on watching, listening, and rental habits; Classify customers by viewing patterns. Predictive policing: Analyze data on past crimes - location, date, time, day of week, type of crime, and related data. Just-in-Time Medical Reporting: EX- Injection notification services: Software analyzes patient’s records, if injections needed, recommends as exam progresses; Blurry edge of medical ethics.
Q2: What are the three primary activities in the BI process?
The four fundamental categories of BI analysis are reporting, data mining, BigData, and knowledge management. Push publishing delivers business intelligence to users without any request from the users; the BI results are delivered according to a schedule or as a result of an event or particular data condition. Pull publishing requires the user to request BI results. Using Business Intelligence to Find Candidate Parts at Falcon Security: Identify parts that might qualify: Provided by vendors who make part design files available for sale, Purchased by larger customers, Frequently ordered parts, and Ordered in small quantities; Used part weight and price surrogates for simplicity.
Acquire Data: Extracted Order Data: Query: Sales (Customer Name, Contact, Title, Bill Yr., # orders, Units, Revenue, Source, Part #), Part (Part, Shipping Weight, Vendor). IS department extracted the data.Actually wouldn’t need all of the data columns in the Sales table. Data was divided into different billing years, which wouldn’t affect analysis. Analyze: First step was to combine the data in the two tables into a single table that contained both the sales and part data. Creating a Customer Summary Query sums the revenue, units, and average price for each customer. Publish Results: Qualifying Parts Query Results- Publish results is the last activity in the BI process. Publish Results: Sales History for Selected Parts: Importance of the human component of an IS. Business intelligence is only as intelligent as the people creating it.
Q3: How do organizations use data warehouses and data marts to acquire data?
Functions of a data warehouse: (a facility for managing an organization’s BI data) Obtain data from operational, internal and external databases, Cleanse data, Organize and relate data, Catalog data using metadata. Components: Programs read operational and other data and extract, clean, and prepare that data for BI processing. An organization might use Oracle for its operational processing, but use SQL Server for its data warehouse. Other organizations use SQL Server for operational processing, but use DBMSs from statistical package vendors such as SAS or SPSS in the data warehouse. Purchase of data about other organizations is not unusual or particularly concerning from a privacy standpoint. Some companies choose to buy personal, consumer data from data vendors Examples of Purchasable Consumer Data: Name, Address, Phone, Age, Gender, Ethnicity, Religion, Income, Education, Voter registration, Home ownership, Vehicles, Magazine subscriptions, Hobbies, Catalog orders, Marital Status, life stage, Height, Weight, hair and eye color, Spouse name, birth date, Children’s names and birth dates. Possible Problems with Source Data: Dirty data, missing values, inconsistent data, Data not integrated, wrong granularity: too fine/not fine enough, Too much data: too many attributes/data points. Data Warehouses V.s. Data Marts: The data analysts who work with a data warehouse are experts at data management, data cleaning, data transformation, data relationships, and the like. However, they are not usually experts in a given business function. A data mart is a subset of a data warehouse. A date mart addresses a particular component or functional area of the business
Q4: How do organizations use reporting applications?
Reporting application: a BI application that inputs data from one or more sources and applies reporting operations to that data to produce business intelligence. Create meaningful information from disparate data sources. Deliver information to user on time. Basic operations: Sorting, Filtering, Grouping, Calculating, and Formatting. RFM Analysis: considers how recently (R) a customer has ordered, how frequently (F) a customer ordered, and how much money (M) the customer has spent.
RFM Analysis Classification Scheme: To produce an RFM score, a program sorts customer purchase records by date of most recent (R) purchase, divides sorts into quintiles, and gives customers a score of 1 to 5. Process is repeated for Frequently and Money. (Top 20%, Mid 20%, Bottom 20%)
Q5: How do organizations use data mining applications?
Source disciplines: Statistics/mathematics, Huge databases, Cheap computer Processing and storage Artificial Intelligence machine learning, Data management tech., Sophisticated marketing, finance, and other business professionals. Sometimes people use the term knowledge discovery in databases (KDD) as a synonym for data mining.There are many interesting and rewarding careers for business professionals who are knowledgeable about data mining techniques. Unsupervised Data Mining: Not a priori hypothesis or model. Findings obtained solely by data analysis. Hypothesized model created to explain patterns found. Example: Cluster analysis: Statistical technique to identify groups of entities with similar characteristics; used to find groups of similar customers from customer order and demographic data. Supervised Data Mining: Uses a priori model. Prediction, such as regression analysis. Ex: Cell Phone Weekend Minutes = (12 + (17.5*Customer Age) + (23.7*Number Months of Account) = 12 + 17.5*21 + 23.7*6 = 521.7 minutes. Market-Basket Analysis: Identify sales patterns in large volumes of data, Identify what products customers tend to buy together, Computes probabilities of purchases, and Identify cross-selling opportunities. Customers who bought fins also bought a mask.Decision Trees: used to select attributes most useful for classifying entities. Unsupervised data mining technique, Hierarchical arrangement of criteria to predict a value or classification, Basic idea- Select attributes most useful for classifying “pure groups.” Creates decision rules. Decision Rules for Accepting or Rejecting Offer to Purchase Loans: If percent past due is less than 50 percent, then accept loan. If percent past due is greater than 50 percent and, If Credit Score is greater than 572.6 and, If Current LTV is less than .94, then accept loan. Otherwise, reject loan.
Q6: How do organizations use Big Data applications?
Huge volume – petabyte and larger. Rapid velocity – generated rapidly. Great variety   
- Structured data, free-form text, log files, graphics, audio, and video. MapReduce Processing Summary: Technique for harnessing power of thousands of computers working in parallel. Big Data collection is broken into pieces, and hundreds or thousands of independent processors search these pieces for something of interest.Hadoop: Open-source program supported by Apache Foundation2. Manages thousands of computers. Implements MapReduce- Written in Java.Amazon.com supports Hadoop as part of EC3 cloud. Query language entitled Pig (platform for large dataset analysis): Easy to master, Extensible, Automatically optimizes queries on map-reduce level. Experts are required to use it; you may be involved, however, in planning a Big Data study or in interpreting results.
Q7: What is the role of knowledge management systems?
Knowledge Management (KM): Creating value from intellectual capital and sharing knowledge with those who need that capital. Preserving organizational memory: Capturing and storing lessons learned and best practices of key employees. Scope of KM same as SM in hyper-social organizations. Benefits: Improve process quality, Increase team strength. Goal: Enable employees to use organization’s collective knowledge. What Are Expert Systems? : Expert systems are rule-based systems that encode human knowledge as If/Then rules. Expert systems shells – programs that process a set of rules. Drawbacks: 1) Difficult and expensive to develop: Labor intensive, Ties up domain experts. 2) Difficult to maintain: Changes cause unpredictable outcomes, Constantly need expensive changes. 3) Don’t live up to expectations: Can’t duplicate diagnostic abilities of humans What are Content Management System (CMS)? : (huge, complex) Support management and delivery of documents, other expressions of employee knowledge. Challenges of Content Management: Huge databases, Dynamic content, Documents refer to one another, Perishable contents, In many languages. What are CMS Application Alternatives? : In-house custom development: Customer support develops in-house database applications to track customer problems. Off-the-shelf: Horizontal market products (SharePoint). Vertical market applications. Ex: Horizontal market: An accounting firm, for example, may license a vertical market application to manage document flow for the processing of tax returns or the management of audit documents. Public search engine: Google, Bing. How Do Hyper-Social Organizations Manage Knowledge? : Hyper-social knowledge management:Social media, and related applications, for management and delivery of organizational knowledge resources. Hyper-organization theory: Framework for understanding KM. Focus shifts from knowledge and content to fostering authentic relationships among knowledge creators and users. Rich directory: an employee directory that includes not only the standard name, email, phone, and address but also organizational structure and expertise. Particularly useful in large organizations where people with particular expertise are unknown. Resistance to knowledge Sharing: Employees reluctant to exhibit their ignorance. Employee competition. Remedy: Strong management endorsement, Strong positive feedback, “Nothing wrong with praise or cash . . . especially cash.”
Q8: What are the alternatives for publishing BI?
Server                                 Report Type            Push Option            Skill Level Needed
Email/collaboration tool     Static                         Manual                                 Low
Web server                        Static/Dynamic           Alert/RSS               Low = static/High = dynamic
SharePo                         Static/Dynamic       Alert/RSS, Workflow     Low = static/High = dynamic
By Server                            Dynam              Alert/RSS, Subscription              High
Q9: 2026?
Exponentially more information about customers, better data mining techniques. Companies buy and sell your purchasing habits and psyche. Singularity: Computer systems adapt and create their own software without human assistance. Machines will possess and create information for themselves.
Three things I learn:
1. An organization can encourage knowledge sharing through a strong management who endorses it and provides positive feedback to its employees.
2. Progressive organizations actually encourages their employees to Tweet, post on Facebook or other social media sites, write blogs, and post videos on YouTube, etc.
3. Big Data is characterized by volumes, velocities, and variations, which exceeds farther than those of traditional reporting and data mining.

No comments:

Post a Comment