Predictive Analytics

What are the five major benefits of data analytics for businesses?

1. Achieved competitive advantage 2. New revenue opportunities 3. Increased profitability 4. Improved Customer Service 5. Realized cost savings from operational efficiencies

What are different predictive analytics techniques? Explain each briefly.

1. Classification: predicting an item class, "Decision Tree", most popular 2. Clustering: finding natural groups or clusters in data 3. Association: what occurs together, "Market Basket" 4. Deviation Detection: finding changes and outliers 5. Estimation and Time Series: predicting a continuous value 6. Link Analysis: finding relationships 7. Web and Text Mining: extracting information from unstructured data

What are the three common mistakes that can ruin predictive analytics efforts in business? Briefly explain each.

1. Don't fall for 'buzzwords'. "Data Science" is simply a buzzword - not a business objective 2. Don't Lead with Software Selection. Prepare staff to manage machine learning integration as an enterprise endeavor, then allow staff a more informed choice of analytics software later. 3. Don't leap to the number crunching - strategically plan the deployment. Know what business problem you're solving for before gathering and analyzing data

What are the most popular barriers to using business analytics?

1. Lack of Resources 2. Lack of awareness

What are the business areas and business questions that can be addressed using predictive analytics?

1. Like Love Procreate 2. Buy, cancel, or quit 3. Default on payment, crash your car 4. Quit Job 5. Vote 6. Commit crime 7. Get sick, die 8. Think, lie

What are the two business areas in which business analytics is most useful?

1. Marketing 2. Operations

What is the problem known as customer attrition for financial institutions? How predictive analytics can be used to address the problem?

???

What are the uses of association rules analysis in predicting customer behavior?

Association is often used for Market Basket Analysis to discover products that are frequently bought by the same customer.

What is the process of building and using predictive analytics models?

Building = data --> predictive modeling --> Predictive Model Using = Characteristics of an individual -->Predictive model --> Predictive Score

What is cluster analysis? What is it useful for?

Finding natural groups or clusters in data. Clustering models are useful for grouping related items together, such as customer with similar demographics

What is the over-fitting trap when using data analytics for decision making? How would you guard against it

Occurs when statistical analysis describes 'random noise' rather than the underlying relationship you need to capture. Keep analysis simple. Randomly devide data set into 'training set' and 'validation set'

What is the confirmation trap when using data analytics for decision making? How would you guard against it?

People only show bosses what they want to see. Seek out opposite skeptic opinions.

What is social network analysis? What is it useful for?

Social Network Analysis is used to explore the relationship between items that interact in networks.

What are the support measure and the confidence measure in association rule analysis?

Support determines how often a rule is applicable to a given data set. Confidence determines how frequently items in Y appear in transactions that contain X

What is predictive analytics?

The application of mathematical computation and models to generate forward-looking insights that can be used to optimize business and IT processes and decisions.

What is the overconfidence trap when using data analytics for decision making? How would you guard against it?

We tend to assume that the accuracy or the probability of success is more favorable than the data suggests. Describe the "perfect experiment" Perform "pre-mortem" by asking subject matter experts to imagine possible failures.

What is association rules analysis technique? What is it useful for

What occurs together. "Market Baskets"

What is the difference between classification and regression techniques? What are they useful for?

classification is about predicting a label regression is about predicting a quantity.