Fluent Essays

You are asked to design a relational database of customer information. Some customer information will not change frequently (e.g. Gender) and other fields change more frequently (e.g. date of last purchase). Which of the following design considerations should be avoided? Question 1 options: Maintain all information in a single table with each row containing all the information about each customer All of these None of these Maintain separate tables that relate to each other with a key, where the tables are arranged by the types of business information (product, demographics, etc.) Maintain separate tables that relate to each other with a key, where tables that contain frequently updated information are separated from tables that are less frequently updated. Question 2 (2 points)   From the textbook chapter, match coding is a database process referring to: Question 2 options: A way to determine what teams play each other in a tournament All of these A way to match interested people with others of similar interests None of these A way to authenticate that a person accessing a database has a correct login and password. A way to assign a unique code to a record (e.g. customer) in the database Question 3 (2 points)   A characteristic of data processing on a BigData platform is that data processing can occur simultaneously. What is the acronym used to describe this capability? Question 3 options: BFD ROI None of these All of these MPP Question 4 (2 points)   From the discussion about BigData paradigms, a computing node is analogous to: Question 4 options: A network junction that routes and distributes requests for fastest response times A numeric address that identifies a consumer on the network One of the many individuals working on calculating an average from the subset of numbers None of these All of these Question 5 (2 points)   You want to test consumer reaction to a product. In terms of random samples and representative samples, which of the following is the best way to think about these? Question 5 options: Neither are important and neither needs to be considered It is more important to have a representative sample than a random sample Both are important and both need to be considered It is more important to have a random sample than a representative sample. Question 6 (2 points)   Imagine that you are looking at a scatterplot with a best-fit regression line showing the relationship between dollars spent (Y) and Age of consumer (X). You notice that the regression line slopes down and that many (most) of the points are relatively close to the regression line. Which of the following conclusions would be true: Question 6 options: All of these The R-squared value should be low because Age is explaining only a little of the variation in Dollars Spent There is no relationship between a scatterplot of the data and the calculation of an R-squared value The R-squared value should be high because most of the variance in Dollars Spent can be explained by Age. None of these Question 7 (2 points)   In the scatterplot with a best-fit regression line showing the relationship between dollars spent (Y) and Age of consumer (X), if the regression line slopes down and many (most) of the points are relatively close to the regression line. Which of the following conclusions would be true: Question 7 options: None of these The correlation would be low and positive All of these The correlation would be high and positive The correlation would be high and negative The correlation would be low and negative Question 8 (2 points)   Imagine that you predict Dollars Spent by including an additional predictor variable, Salary (as well as keeping Age in the model). All coefficients are negative and statistically significant as is the R-square. Which of the following statements are accurate? Question 8 options: None of these Each predictor is negatively correlated with Dollars Spent Consumers that are older and that have higher salaries spend less on purchases If we know a new consumers Age and Salary, and their Age and Salary are within the range of values that the model was built upon, we can predict how much dollars the consumer will spend. All of these Question 9 (2 points)   You have a set of data points on Millennials (age) and their predicted Customer Satisfaction scores that youve analyzed and modeled with regression. You now want to draw conclusions about Boomers based on the same data and model. This would be an example of ... Question 9 options: Interpolation and is not a good idea Extrapolation and is a good idea. All of these Extrapolation and is not a good idea None of these Interpolation and is a good idea Question 10 (2 points)   If there is correlation between Customer Satisfaction (Y), and Amount Spent (X), of r=0.50, then a simple regression model using amount spent to predict customer satisfaction will have how much explanatory power (as a percentage)? Question 10 options: 100\% 25\% 50\% None of these Any of these Question 11 (2 points)   Which of the following describes a key difference between a CART vs a CHAID decision tree methodology for building segments? Question 11 options: CART results in shallower or bushier trees CART makes binary splits at each branch CHAID trees cannot make a split on both a variable, and value(s) of the variable CHAID splits are determined by the analyst, and not by the underlying algorithm in the software All of these None of these Question 12 (2 points)   Regardless of the segmentation method used, a segmentation analysis addresses which of the following questions: Question 12 options: How many consumers will be in the segment(s)? What is the expected response rate of the segment(s) How much revenue or products can I expect from the segment(s) All of these None of these Question 13 (2 points)   Imagine that you run a Marketing Analytics team. Your team has run a Segmentation Analysis and all 10,000 customers in the company database have been assigned to one of three segments: Best, Good, Not So Good. Your Chief Marketing Officer and Chief Financial Officer have asked how much revenue might be expected if the company presented an offer only to customers in the Best segment, but each customer is limited to one item each item brings in $5. Your analyst has calculated an expected revenue amount of $50,000. What might you conclude? Question 13 options: $50,000 would require a response rate that would not be possible $50,000 would mean that the company would have to include customers that are not in the Best segment None of these - your analyst could be right, so present the $50,000 answer All of these - so have the analyst resubmit the analysis Your analyst has made a critical calculation error. Question 14 (2 points)   According to the Economics Cohorts segmentation chart, what segment code would a single (unmarried), 31 year old, living and working as a marketing manager in NYC, and earning $85,000 belong to? Question 14 options: A6 C17 E27 All of these None of these Question 15 (3 points)   You are performing a Segmentation Analysis and want to target customers that are most likely to respond. A cell in your analytic sample has a response rate of 4\%, and the overall response rate for all cells is 1.5\% If your targeted group must have a response rate index greater than 2.0, would this cell be part of the Target group - Yes or No, and explain why. Question 15 options: Question 16 (2 points)   You’re an expert in recommending marketing strategies based on Lift calculations. You have a database table showing groups of customers in four quartiles based on scores that predicts likelihood to respond. Incremental and Cumulative Lift values for each quartile are available. Your client wants to launch a campaign to send an offer to customers. The campaign has to have at least a 6.5\% overall response rate. To determine if the campaign should include names that are in the 2nd quartile group you would need to check Question 16 options: The overall RR across all quartiles to see if it at least 6.5\% None of these All of these The incremental RR of the 2nd quartile to see if it is at least 6.5\% The cumulative RR of the 2nd quartile to see if it is at least 6.5\% Question 17 (2 points)   Imagine that you plot a cumulative gains chart from a model, and the curve for the model closely follows a linear relationship showing that by targeting ‘x’\% of the responders, you can capture ‘y’\% of all orders, and the model’s line is such that X=Y. You can conclude that: Question 17 options: The model performs better than not having a model All of these None of these The model performs about the same as not having a model The model performs worse than not having a model Question 18 (10 points)   You’re an expert in Customer Service Analytics, and Finance. The CFO is concerned that the company will lose money because of the company’s generous return policy. Customers get free shipping on purchases and within 30 days of a purchase, customers can return an item they purchased, get a full refund, and your company will even pay for the shipping costs for customers to return the item. You collected some data and find that the average price of an item is $200. There’s a 20\% chance a customer will return it for a refund within 30 days. Shipping costs average $10 each way. Under these circumstances, calculate the long term Expected Monetary Value (EMV) for the company. Frame up the problem in terms of the financials, the probabilities, and show the EMV equation and numbers in your calculation. Question 18 options: Question 19 (3 points)   The CEO says hes heard you did some analysis on the companys return policy. He doesnt think hell understand the equations so he wants you to explain to him what your EMV result means for the company and whether the policy is a good idea or not. How would you respond. Question 19 options: Question 20 (2 points)   Understanding the Breakeven point in marketing analytics is important because: Question 20 options: It can help manage a business response to any changes in Costs, Profits or Response Rates It quantitatively relates Profit, Costs, Response Rates None of these It represents the point that a business needs to exceed to be successful All of these Question 21 (2 points)   Based on the text, which of the following is an example that illustrated why an NPV analyses is important? Question 21 options: It provides a way to assess future what-if effects under various business operating conditions None of these It provides insights to current reasons for customer satisfaction If provides a historical look at past performance of the business All of these Question 22 (2 points)   You have calculated LTV values for customers over a period of 3 years, including NPV adjustments for each year. To calculate the average LTV of a customer for Year 3, which of the following apply? Question 22 options: All of these Retention rate is usually assumed to be a fixed value of 15\% None of these The exponent in the NPV calculation is = 3 The denominator of the average LTV calculation is the number of customers that you started with Question 23 (2 points)   A hypothesis test with an alpha level = 0.05 is equivalent to a 95\% confidence interval analysis Question 23 options: True False Question 24 (2 points)   Generally, as the confidence level percentage increases, the interval between the lower and upper bounds ___________ Question 24 options: Increases Decreases Either decreases or increases Neither - Confidence Interval and Confidence Level are not related. Question 25 (2 points)   When you test a hypothesis, you are essentially analyzing data to determine if something happened, or nothing happened Question 25 options: The null hypothesis means nothing happened and the alternative hypothesis means something happened The null hypothesis means something happened and the alternative hypothesis means nothing happened Either of these Neither of these Question 26 (3 points)   You have been asked to shift from analyzing Marketing data to analyzing Covid-19 Test data. In terms of COVID testing, describe and differentiate a Type 1 error and a Type 2 error Question 26 options: Question 27 (10 points)   Imagine that you work in web media, and your company publishes news and analysis. Success is measured by the amount of time visitors stay on the site consuming content (longer time is better). Two websites are created and tested: Deep-Analysis (DA) vs. Just-Headlines (JH). An analysis of the duration of time spent by unique visitors is as follows: Deep-Analysis: 1000 unique visitors; Duration= 2:40 (hh:mm); std.deviation = 25 minutes. (or in decimal numbers, mean=2.66, sd=0.42) Just-Headlines: 1000 unique visitors; Duration= 2:15 (hh:mm); std.deviation =30 minutes. (or in decimal numbers, SLS mean=2.25, sd=0.50) Test the null hypothesis claim at 90\% confidence (or alpha=.10) that the difference in time spent between these two sites is not significantly different from zero. (i.e. Note that this means that the null hypothesis is claiming that the same amount of time is spent on these sites). Do you retain or reject the null hypothesis? Question 27 options: Question 28 (3 points)   Based on your analysis of comparing the time visitors spend on the DA and JH websites, which site would be more attractive for advertisers to use? Explain. Question 28 options: Question 29 (3 points)   If you calculate a 90\% confidence interval to estimate the true difference in time spent between the sites, and also calculate a traditional hypothesis test (using an alpha=0.10), should the conclusion based on Confidence Intervals be the same as the conclusion your would reach in the hypothesis test? Explain or demonstrate via calculations. Question 29 options: Question 30 (2 points)   In digital marketing, many marketing touch points may be presented to a consumer (personalized email, targeted banner ads, etc.), and the consumer may decide to make a purchase (even offline) sometime later. The ability to measure, and credit, the fractional effect of each marketing touchpoint to the purchase behavior has been modeled by which of the following approaches? Question 30 options: Regression Models Attribution model All of these Hidden Markov Models None of these Question 31 (2 points)   In the context of digital marketing analytics, which of the following are used to track and/or identify a consumers online behavior. Question 31 options: Cookies, IP Address All of these None of these Device identifier Question 32 (2 points)   The death of the cookie problem has an impact on which of the following marketing analytics capabilities Question 32 options: All of these Tracking of consumers on websites None of these Provides consumers with greater privacy over their web visit behaviors The ability to send personalized and targeted offers to consumers based on their web visit behaviors Question 33 (2 points)   If you were to work on a digital marketing campaign localized to China to effectively reach Chinese consumers, which of the following are important considerations Question 33 options: All of these Connect with target consumers via direct mail None of these Connect with the target population on Google and Facebook as primary platforms. Connect with the target population on Tencent and Alibaba as primary platforms Question 34 (3 points)   Your mom calls you and says that shes been hearing that companies are looking to hire data scientists. She asks you if thats been covered in a slide in that Database class youre taking, and if so, what disciplines combine in Data Science? How would you respond? Question 34 options: Question 35 (2 points)   What best describes what is happening in the following piece of SAS code? proc corr data=nyugrads2021; var salary age; run; Question 35 options: SAS is issuing a correction to SALARY and AGE in the nyugrads2021 dataset SAS is generating a correlation between SALARY and AGE in the nyugrads2021 dataset. All of these None of these SAS is issuing a correction to the dataset name Question 36 (2 points)   What is the correct SQL statement that would be part of an analysis of all records in the nyugrads2021 table? Question 36 options: SELECT allrecs FROM nyugrads2021; All of these SELECT * FROM nyugrads2021 WHERE age >20; SELECT * FROM nyugrads2021; None of these Question 37 (2 points)   Imagine that two different analysts were independently asked to build an RFM-only segmentation model from the same dataset. If both analysts executed their work correctly, their final segments and scores should be identical. Question 37 options: Both of these True - they should come up with matching answers if theyre using the same data and do this correctly False - there are many ways to combine and use R-F-M variables, so theres no guarantee that their answers would match Neither of these Question 38 (3 points)   In a discussion about emerging trends in databases someone asks whether Translytical platforms represent a movement towards one platform that can do many things and not a move towards more specialized platforms being dedicated for very specific purposes. Do you agree? How would you respond? Question 38 options: Question 39 (2 points)   One of the emerging trends is a difference between a Data Warehouse vs. a Data Lake. Which of the following describes the difference between a warehouse and a lake? Question 39 options: In terms of the store analogy, a Data Lake can take in many different product-types and only sorts and organizes them at check-out or when needed. A Warehouse has products pre-sorted and organized ahead of time and before theyre needed. In terms of the store analogy, a data lake can only store products that are related to liquids and fluids. None of these In terms of the store analogy, a warehouse is bigger and more flexible than a data lake about what types of products can be stored All of these Question 40 (2 points)   Please rate this exam (Free points, as any answer is correct) Question 40 options: Much easier than expected Easier than expected About what I expected More difficult than expected Much more difficult than expected