Transcript Mac: new
Introduction to Management Information Systems Chapter 9 Business Intelligence and Knowledge Management HTM 304 Fall 07 Business Intelligence System Chapter 7 & 8: Operational data and information. Information Flow designed to facilitate corporate daily operation Tracking orders, inventories, and shipments Managing account receivables, payables Storing employee information, addresses, HR benefits Chapter 9: Systems that takes daily operational data as input and produce higher level “business intelligence” Analyzing order patterns, data relationships, clusters for strategic planning and forecasting Analyzing customer relationships, identifying potential business problems and business opportunities GPS for citation appeal? 2 Business Problem Carbon Creek Gardens Mary Keeling retails trees, plants, flowers, soil, fertilizer, etc. Ran into a good customer – hasn’t shopped in a year Salesperson was rude Mary realizes she needs better information 3 Challenge of Data Analysis Data Volume Facts: Study at UC-Berkeley: Total of 403 petabytes new data created in 2002 403 petabytes = all printed material ever written Printed collection of Library of Congress = 0.01 petagytes 400 petabytes ~ Collection of 40,000 Library (size of LOC) Directly related to Moore’s Law Today, storage nearly unlimited Drowning in data & starving for information! 4 2.5 Exabyte by 2007! 5 Business Intelligence Tools BI tools: search data to find patterns or information Reporting tools: Read and process data, produce and deliver reports Used primarily for assessing the past and current situation Data-Mining tools: Process data using sophisticated statistical techniques Searching for patterns and relationships among data In more cases, used to predict (give probabilities of loan default, id theft, etc.) Differences of reporting and data-mining tools Reporting tools use simple operations like sorting, group, and summing to provide description of existing data (mainly descriptive statistics) Data-mining tools use sophisticated techniques (including inferential statistics) 6 BI Systems BI System: The IS that incorporates BI tools Purpose: to provide the right information, to the right user, at the right time. Help user accomplish goals and objectives by producing insights that lead to actions 7 Two types of BI systems Reporting System Use reporting tool to produce status report: generate report showing customer cancelled important order Deliver the report to the right person at the right time: alerts salesperson with bad customer news in time to try to alter the customer’s decision Data-Mining System Use data-mining tool to predict the events and probabilities: Create equation to compute the probability that customer will default on loan Deliver the probability to the right person at the right time: Use 8 equation to enable bankers to assess new loan applicants Reporting System Purpose: To create meaningful information from disparate data sources and deliver information to proper user on timely basis. Reporting system normally generate information from data through 4 operations Filtering Sorting Grouping Making simple calculations 9 Example: From Data to Report 10 Example of Online Report Systems 11 Components of a reporting system 12 Report Mode Push report Organizations send push report to users according to preset schedule Users receive report automatically Pull report Requested by user User goes to Web portal or digital dashboard and clicks button to have reporting system produce and deliver report 13 One Solution to Carbon Creek Gardens RFM Analysis report: analyzing and ranking customers according to purchasing patterns Simple technique considers how -- how recently (R) customer ordered -- how frequently (F) customer orders -- how much money (M) customer spends per order 14 R F M RFM Analysis To produce RFM score, program first sorts customer purchase records by date of most recent (R) purchase Divides customers into five groups and scores customers 1-5 Top 20% of recent orders given R score 1 (highest) Re-sorts customers on order frequency Top 20% of most frequent given F score of 1 (highest) Sorts customers according to amount spent 20% of biggest spenders given M score of 1 (highest) 15 Example of RFM Analysis output Exercise: Who should be your major marketing force target? Write down your analysis to explain why. 16 Data Warehouses & Data Marts Data Warehouses and Data Marts: Prepare, store & manage data for data mining and other analyses Report systems report up-to-date status information Cumulative reports stored in warehouse can be used for further analysis. multi-dimensional “data cube” East Central West 50 40 90 50 40 90 50 40 90 60 60 120 60 60 120 60 60 120 100 80 140 100 80 140 100 80 140 2005 2004 2003 Nuts Screws Bolts 17 Data-Mining Systems Application of Statistical Techniques to find patterns and relationships among data to classify and predict. Represents a convergence of Disciplines Statistics Mathematics Artificial Intelligence Machine-learning fields in Computer Science 18 Example of Data Mining Customer Analysis Group 1 – average age 33, owns at least 1 laptop, 1 PDA, drives high-end SUV, buys expensive children’s playing equipment Group 2 – average age 64, owns vacation property, plays golf, buys expensive wines and designer children’s clothing ID Theft Risk: good credit rating live in San Diego outstanding home loan mortgage rarely travel, grocery shopping weekend, weekly gas refill Alert? When Hotel check-in at Las Vegas? 19 Buying LV handbag in Miami?