ISM 3541 EXAM 1/KERWIN (LATEST) Download to Score A: Florida State University.

Module 1: Data analytic thinking Introduction - Learn! Be a constant learner - Let’s data tell your story - Helps you map out how to get to a decision - Where are you going? Do you have a plan for it? How does data help with those decisions? - Data driven decision making - Big picture is to do with thinking big!! - Found a concrete experience to relate to Object oriented thinking - In collaboration with business stakeholders, data scientists decompose a business problem in subtasks o The solutions to the subtasks can the n be composed to solve the overall problem o Some of these subtasks are unique to the particular business problem, but others ae common data mining tasks - Noun: person, place, thing (object) o Objects have properties and methods - Properties are adjectives (words that describe a noun) - Adjective: Words (properties) that describe a noun (object) - Verb: actions (methods) that can be invoked by the noun (object) - What are some properties/ attributes of each data -driven BUSINESS decision -making PROBLEM? o What is your proc ess for making a decision?  Do you need data?  What kind of data do you need?  How do you acquire that data? o What are the properties of a business problem?  Probably money involved  Probably stakeholders involved  Probably risk involved o What are the methods  Acquire resources  Find a solution  Maybe find people and technology to enable this solution VORTEX / decision making - What is the process you go through when making a decision? (think about destination) o Sometimes trying to solve from different angles - Vort ex Model – analytics will tie these 3 views together, tells the story of each of these levels, and get them to align o Bird’s eye view (30,000) feet  Example: flying over your hometown and being able to possibly point things out to someone who has never been there before because I have been at the ground level and I have been at the tree level POV’s  If you’re in the plane you can tell there’s the stadium but you can’t tell you the score because its 30,000 feet above and you can’t see how many people are  When t alking about organizations – executives live at 30,000 feet view  Has a great view of how people are doing at the tree and ground level and can give them good feedback so that they get to the destination they’re trying to get to o Tree level view  Basically m anagers  Have multiple people they’re responsible for  Need a vantage point so they can see them o Ground level view  “Worker bees” o Example of the vortex model: Sports team  Owner of the team – Bird’s eye  Coach – bounce between bird’s eye and tree level  Enough couches – the regular coach could be at the bird’s eye and he makes the other coaches look at tree view  Team – ground level o Need communication! Data mining tasks – common methods - What are some properties/ attributes of each data -driven business de cision -making problem? o Properties – adjectives  Descriptors of the object  Properties – describe the object business problem o Attributes – what we’re calling the elements that describe the data set - Object: Business problem o Attributes  Unique  Goals  Stated obje ctive  Desires  A want, a need  Constraints  Constrained by budget, etc.  Personalities  Because people are involved there will be personalities o Methods  Decompose problem into subtasks  Pattern recognition (avoid recreating the wheel)  Invest focus on interesting parts of the process requiring human intelligence  Invoke creativity o Common data mining tasks (methods extended )  Classification  Predicts whether it will happen  Regression  Predicts how much it will happen  Similarity matching  Match similar individuals based on data known about them  Clustering  Group individuals and population by their similarities  Do our customers form natural groups and segments or not>  Co-occurrence grouping  Find associations between entities based on transactio ns involving them  Ex. What items are commonly purchased together o They put certain products next to each other  Profiling  Attempts to characterize typical behavior of individual , population, group, etc.  Ex. What is the typical cell phone use of this customer segment?  Link prediction  Attempts to predict connections between data items  Common in social networking systems  Ex. You and Perry have common friends  Data reduction  Large data set and then taking a sample size and do some mining  Casual modeling  Helps us u nderstand what events and actions influence others  Ex. Did the advertising increase the selling of products? Data mining tasks – supervised vs. unsupervised methods - Do our customers naturally fall into different groups? o No specific purpose or target for this specific question  When there is no specific purpose or target the data, mining is referred to as UNSUPERVISED - Can we find groups of customers that have particularly high likelihood of cancelling their service soon after their contracts expire? o This has a specific target defined (will customer leave when contract expires)  This is an example of a SUPERVISED method - UNSUPERVISED