Module 5 Flashcards
What are some example use cased for predictive analytics?
o Critical decision support
• Risk of patient developing certain conditions
o Cross-Selling
• What is the customer who bought x likely to buy next
o Fraud Detection
• Fraudulent spend patterns with credit cards
o Customer Retention
• What are behaviors of customers that left?
• Big retailers have customer cards to see
o Risk Management
• Capital asset pricing model predicts optimal portfolios
• Underwriting and Scoring
Which advantages does in memory BI provide?
o Improved performance
• RAM> Disk
• Optimized memory usage- columnar compression
o Enables analytics that were not possible before
• Highly explorative/ interactive analytics
o Reduced development cycles- faster project delivery- Key reason
• Less time spent on performance optimization
• Generally less objects needed- data marts, aggregates
• Reduced data redundancy
• Errors are faster to find for IT professionals
o In Essence
• More hardware
• Less manpower
What are its disadvantages & risks?
o Disk is persistent, RAM is not- All data in RAM and there’s a power outage it’s all gone
o Still comparably new technology
o Less robust than traditional technology
o Complicated recovery
o Polypol not fully developed- few providers and high prices
o Specialized hardware needed- High in memory, massive parallel computing, however significant investment needed
o Not many knowledgeable people on the market
o Highly oversold- manage expectations
• Does not fix the data quality problems
o High usage and increasing number of users causes bigger memory needs
o Leads to lack of reflection on data model/ architecture
• They know it will be fast, so they develop quick and dirty
Knowing the pro ́s and con ́s when would you recommend BI on
demand?
o Knowing the pro´s and con´s when would you recommend BI on demand/cloud?
• Small company or departmental solution
• Data integration is less important
• Cloud provider delivers missing IT skills
• Temporary/ exploratory BI system
• Proof of Concept
• One time analysis
• Specialized cloud software
• That provides unique functionality
• Comes with already needed data integration- Huge benefit for programmers or IT people/ whatever they’re called
• Gravity of Data
• Analyze external data
• Makes sense to store data where it already resides
• Data of limited confidentiality
Why does it make sense to have a BI strategy within a company?
o Some quote about treating BI as a key corporate asset- with all the funding and oversight it needs or else you will fall behind your competitors
o Why does it make sense to have a BI strategy within a company?
• Demonstrate value add to company strategic objectives
o BI is a highly complex topic requiring
• Alignment of many stakeholders
• Fit for purpose organizational model
• Senior MGMT sponsorship
• Long term vision and clear decisions
• Many technologies to select from
• Balance governance and flexibility
• Open vs controlled access to information
How do you figure out if a new trend emerges that might be
relevant to your business?
o Market Research Providers
• Gartner
• Forrester
• Etc
• Pros- They deliver quality at a price, credibility, experts
• Cons- They sell to everybody, not getting competitive advantage. At a high level, not very specific so they can sell to everyone
o Attend Conferences
• Cons- Pricey and the speakers can be bad
o IT Consultancy
• Pros- Specific solution
• Cons- Also sell products like IT solutions, SAP will try and sell you their solution. Look for independent consultant
o Talk to peers and colleagues
• In your industry
• In other industries
• Pros- He claims it usually works well, usually get quality information
• Cons- Competitors may try and mislead you and try to steal information from you. This people are like you… not smarter or dumb (speak for yourself simon)
