Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis? Don’t use an example we’ve covered in class.
What advice would you give someone regarding how to select the right predictor variables for a decision tree analysis?
An example of a business question that you could try to answer by using a decision tree could be, “How do customers find our business’ website?”. A decision tree could help find out how likely people find the website through an organic search or by clicking on an ad, to better understand the traffic that comes to the website. This could also help the business change their marketing strategies to reach more people through different ways. The data that would need to be collected will have to do with what percentage of people find the website through different ways as well as defining what those specific ways are (i.e. Google search, ad clicks, etc.). The advice I would give someone regarding how to select the right predictor variables for a decision tree analysis would be to make sure you look at all the traffic coming to your website so you have a large, representative sample size that is very accurate to which means of traffic is most popular.
A decision tree could really help with an airline with deciding the profitability of different flight routes, at different times of days, with different sized planes. The biggest difference is the route, and how much demand there is for that route. Then you could split that up by the time of day and the demand for flying that flight at that time of day. Then you would split it up finally by the different sized profits the different sized planes could bring in. This would help an airline create a flight schedule that would maximize profitability by matching demand for each node.
One business question a decision tree can answer is whether someone will book first class flights. The data we should collect to perform this analysis includes the age of people, their annual incomes, genders and jobs. When selecting the predictor variables, we should think about if the predictor variables are relevant to the problems or if they contribute enough to differentiate the outcomes. We can use the Chi-square test to check if the predictor variable is important enough to create different groups.
A business decision that could be answered with a decision tree is determining whether a business should make an investment or not. The expected loss/gain for the investment depending on certain circumstances (good or bad economy, amount of investment) can be determined with a decision tree. This can help a business determine the probability of an investment or expansion being worthwhile. If the expansion is profitable when the economy is good or bad, the company should invest. If it is not profitable no matter the circumstances, investment should be held off. My advice for setting predictor variables would be to set the most important variables early on in the tree, try and have a large sample size of the predictors to see how likely they are to happen, and to find accurate probabilities of the various circumstances.
A business question you could answer using a decision tree is: “How can I predict whether a student will apply to Temple”? This question would help Temple University determine what it should be doing more/less of to effectively reach its target demographic. The following are examples of data that can be collected to perform the analysis: Student GPA, # of Student’s Extracurriculars, # of Student’s AP Courses, etc; essentially any data that describes a student. The advice I would give someone regarding how to select the right predictor variables for the decision tree analysis would be to choose predictors based on whether they maximize distinct outcomes and separate outcomes. I would also advise them to use a substantial dataset (ex: all of last year’s incoming freshman applicants) for data collection.
A business question you can answer with a decision tree is: “Should we open a new restaurant in this location?”. This question would help a restaurant owner who has a few locations, but is looking to expand to add more stores. Some of the data I would collect is .the population of the zip code being considered. This would be a good indicator of potential customers. Another piece of data I would collect is the number of restaurants currently in business in the location. Household income and number of people living in each housing unit. The advice I would give to this business is to choose a location with a high population so that there are customers, a high household income so that there are people willing to spend, and a zip code with a lot of families because busy parents will want to take their families out to eat.
The business question to be answered is whether a college student will play volleyball based on any of the following combination of variables: Age, Gender, GPA, number of courses taken in the current semester and fitness status (active = 1 or not active = 0). A decision tree will provide a simple illustration of branching and splitting of these predictor variables to determine the likelihood of a college student playing volleyball. When choosing the predictor variables, be sure select variables that are relevant to the analysis (for example, do not use the student’s TUID as a predictor variable). It is also important to select variables that give insight to the outcome (not every variable may be a useful predictor variable to explain the outcome of a business question).
A business decision tree example: For a auto insurance company, whether they should offer a person insurance plan. This question helps the company manage business risks. Variables in this tree includes the person’s age, gender, occupation, license’s history/length, driving score/experience, etc. The insurance company will determine whether to make this person become their member. Also, this tree can further help the company to calculate the premium of an individual. Depending on how the customer’s background is, the system will set different levels of premiums.
One business decision that could use a decision tree would be recruiters looking to hire new graduates. The data they would collect would be GPA, school, and major. Although the decision tree would be extremely big with many different nodes, the system would be able to automate the decision tree process for every candidate making it much easier for the recruiter to sort through all the applicants. If it is for a specific position, the first set of decisions would be for major, but if it was more general, they would maybe want the first set of decisions to be on GPA.
A business question, not discussed in class, that you could answer using a decision tree would be: “how likely is it that someone who buys product A is going to buy product B?” This could help with providing recommendations on e-commerce sites, such at what Amazon does. Data to be collected could be: how much people spend on their initial product, the accompanying product(s) costs, what items are bought together, demographic of people that shop for certain items, time that items are bought and how often the person shops on the site and for particular items. Anything that describes specific buying behaviors of one customer could eventually show how other customers will behave. My advice for choosing predictors would be to use predictors that are more easily quantifiable. Additionally, I would tell them to choose predictors based on the data they collect. Also they should continue collecting large samples of data based on predictors that will show a relationship between consumer behavior and the prediction.
A business question that you could answer using a decision tree is “Should the company book flight or train ticket for a team of employees?” This question will help the company to determine what is the most efficient way and also can save more money for the company. The data should be collected are the prices of flight and train at same time period, the time of flight and train from origins to destination etc. The most important variable is the price. The company should make a budget of how much the maximize price could be acceptable, and then compare this two methods’ time whether it’s appropriate or not. My advice is to set the most important variables at first to filter any other unacceptable choices. And then set relevant variables to make a better comparison.
A business question that can be answered using a decision tree is “How can I predict how successful a 7-Eleven will be in a certain location”. This question would help 7-Eleven corporate decide on the best locations to expand their convenient stores. Data collected would include population, age ranges, number of similar convenient stores near the location, income, and whether or not they own car. This is a good indicator of potential customers in the area, how competitors could affect their success, if there are households with high incomes that can afford to shop at the convenient stores, and if those individuals can get to the store by car or not.
A business question that can be answered through a decision tree is who is most likely to respond to a promotional offer. You could sort through your customer database with a decision tree to see the most significant predictor and choose to send your promotional offers based off of that to save costs on mailing. Good predictors would be gender, number of purchases, and income. A big tip for decision trees is to make sure you are choosing predictors that are relevant to your business question.
One question that can be answered with a decision tree is the chances of an insured getting into a car accident in a auto insurance company to effectively price the premium based on the risk that each person is assigned to. Information such as age, gender, years of driving, is there involvement in accidents in the past, the type of car, the color of the car, the region that the insured live, etc would all be data needed to perform the analysis.
To select the right predictor, I think it would be important to know what you’re asking for. In this case, it would be the risk (how likely the insured would get into an accident) of this person based on the past data.
A business question not discussed in class that a decision tree could use is “does this individual qualify for our job listing?”. There are many questions that get asked on a job application. By using a decision tree, if would make it much easier for the hiring manager to sort through the data from the applicants. The root node could be application submitted and then the tree could split on major questions such as “legally allowed to work in the US”, “Not convicted of a felony”, “GPA higher than 3.0”, “major”, etc. The nodes would eventually get to the leaf node which would be “invite to interview or don’t invite to interview”.
A business question you can answer with a decision tree is: “Should we build a stadium in this location?”. This question would help a city decide whether or not a new stadium would be good for the city socially and economically. There are many factors involved in this question, many being political, but the best approach starts with gathering data on the proposed area. I would collect data on the population of the area in which the stadium would be built. Another piece of data to be collected would be the number of businesses in the area’s location. I would tell developers to build in a location that needs economic help and a smaller population so those residents aren’t forced to relocate.
A business question that could be answered is determining whether a customer should be able to get the Chase Sapphire Preferred card or American Express Gold card. This question will help credit card companies qualify the right customers. The data to collect for this analysis includes having a credit score above 700, making sure you don’t have too many credit cards open, and your credit utilization. Some advice I would give to pick the right predictor variables include having a large sample size to pick from because it will produce the best outcome on how to pick the right applicants.
A business question one could answer using a decision tree is as follows – does the following candidate qualify for a car lease?
The variables you would see within this tree are: Name, Address, Age, Income, Car Insurance Provider, Accident History, Employment Status (Yes=1 OR No=0).
The decision tree will utilize these predictor variables by “branching” and dividing up these categories in order to determine whether the applicant qualifies for a lease.
Important Note: When choosing variables for the decision tree be sure select variables that are relevant to the analysis (choosing “Race” has nothing to do with applying for a car lease, so it would not be a valid variable. It is important to choose variables that will help predict/determine the outcome to whatever situation you are analyzing. For this example in particular, an applicants “income”, “accident history” and “employment status” would tell us tons about whether this individual is a reliable person to lease a car too.
A business question that you could answer using a decision tree would be what types of applicants would a company accept for a job position. This would help the company to figure out which resumes and applicants to put through to an interview, and which do not fit the qualifications for the position. The data the company would collect to decide this would be college attended, major, GPA, work experience, and leadership experience. If the students GPA is not over a 3.0, they might be automatically thrown out, unless they went to a prestige University or have a strong work experience background. The advice I would give someone in regards to selecting the right predictor variables for the decision tree would be to allow as many applicants as possible in order to get the largest pool of applicants. The more applicants, the greater the chance that the company will get the strongest employees with the qualifications that they are looking for
A business question that can be answered using decision tree is whether an athletic shoe retailer can do to raise sales. To raise sales there are multiple options: increase advertising, cut prices, launch royalty cards, etc. The data to collect for analyzing would be the number of customers after each option being executed, total sales and total costs after each option being executed. A good advice to pick the right predictor for this analysis would be gathering data in the right demographic; for example, teenagers and young adults are the most frequent athletic shoes consumers so gathering data from them can return better results than gathering from older ages who do not consume athletic shoes that much.
A business question for NBA scouts could be “What is the likelihood of a certain player becoming an All Star in the league?” Based on data such as age, height, points per game, assists per game, rebounds per game, and plus/minus rating, a scout could use a decision tree to determine how good a player will be at the professional level. Some advice I would give to pick the right predictor variables is to choose variables that have a direct impact on winning games, as opposed to less important variables such as how well a player played on a certain day of the week compared to others.
A business decision that could be answered with a decision tree is ”which shoe brand people buy most between Nike and Adidas.” By using tree decision tree we can find out the average of the people which shoe brand they wear more. Data should be collected on how many people wear Adidas and Nike shoes, how often they go to store, and how much time they spend on buy shoes. The most important variable is player fans, how many people follow the players who Nike and Adidas. The advice I would give to The advice I would give someone to chose the predictor variables for the decision tree analysis based on the data. The should collect data from the busy location stores, and data of people followers their players, and they should collect large data to get a better result.
A decision tree could help to answer the question “should I buy this used car?”
The variables in this decision tree would include amount of mileage, year of car, make/model, price and what kind of condition it is in. The most important thing when choosing variables for a decision tree is to make sure they directly relate and are relevant to the question that is being asked.
An business question that can be answered using a decision tree would be if a car manufacturer is thinking about what kind of car to launch? if they want sedan,suv,truck,etc… First they would do market research and look for what is most demanded. And also finding out how much people are willing to pay for the model, which can include if the people want more of luxury or just simple car. The advise i would give is to follow with what people want. If every one wants a compact SUV then the best choice would be to choose that option. Also find the accurate number of demand.
A business question that you could answer using a decision tree is the factors that cause consumers to buy more organic options. In order to answer this question, I would need to collect data about the consumer’s monthly household income, the age of customers, the amount of individuals in the household, and the gender of the customer. In order to think of the right predictor factors to use for a decision tree, you have gather data that you believe will indirectly or directly affect the decision tree. For example, age is a big determinant of whether not an individual will buy certain products or engage in specific activities. As a result, age has a positive correlation to differing preferences in buying organic products.
A business question that can be answered using decision tree is which stock is more valuable for investor. For investment, there are multiple options: Stocks, MF, Bonds and etc, all the useful information are collected for analyzing what the optimistic expected value and what the worst expected value. Adding all the expected value up, and compare all the choice, then pick the most valuable one. For example, the stocks growth rate 40% and is the highest in all the choice, and declining rate only 13%. relatively lower than others. under the same probability, most people might choice the stocks as an investment.
A business question that could be answered using a decision tree is “How likely is a customer to subscribe to our e-mail newsletter?” A decision tree could determine which customers are most-inclined to join the newsletter based on a number of different factors. The data I would collect to perform the analysis is gender, age, purchase history, amount spent on products, or any other data that would indicate how loyal a customer is to our brand. In terms of advice on selecting predictor variables, I would encourage them to use a large data set of customers to gain an accurate tree, as well as only using predictor variables that increase accuracy of the tree by a significant amount.
A decision tree can help answer the question “should a business outsource its tech support?” The data that needs to be collected includes number of support tickets generated per month, whether issue requires remote or in-person support, and length it takes for IT team to solve issue. A business that has a large number of tickets that require longer assistance and/or in-person assistance would result in ‘yes, outsource’, My advice is to choose predictors that are statistically different and can further divide and relevant to the question.
A business question that could be answered using a decision tree is “Which track will an advertising student choose?” A decision tree could help a university, such as temple, determine which classes a track a student will decide on in their sophomore year. From this the university can determine how many teachers to hire, classes to run, etc. The data that would be collected would be the gender, age, classes taken previously, grades in classes taken previously, GPA, credit hours completed, which extra curricular clubs they are members of, major. My advice would be to select a large sample size and to choose the correct variables to lower your risk of inaccuracy.
A business decision we can solve by using a decision tree is “Who is performing well enough to receive a promotion?”. A decision tree can help solve this because it will show who is performing the best, who is performing moderately, and who is performing the worst. Some data we would collect for this decision tree would be their time sheets, performance, teamwork, and evidence from other employees. Each of these things can prove that the certain employee can indeed receive a promotion. Advice for people to choose predictors in a business tree is to choose the main ideas that you want to show. Main ideas and important details are a good indicator of picking the predictors.
Business questions that are answered everyday using a decision tree can include “How can we determine an appropriate life insurance policy for a client?”. By use of personal data such as: age, gender, demographic, income, preexisting conditions, medication, stress” a decision tree can aid in creating an individualized policy.
To aid in collecting the correct data sets for a decision tree, I would strongly advise requesting information from the client that is pertinent to their risk of becoming injured or ill. Through taking all personal aspects into account, such as marital status, exercise patterns, dietary routine, a decision tree will effectively paint an accurate picture of your client’s risk.
The business question that could be asked with a decision tree is: What type of a car customer can purchase?
It will determine the factor of levels of income ,age, models of the car and the price. Decision tree can help an individual to make a decision effectively and consider all the options.
A business decision a student could use R to answer would be “what Masters Program should I apply to based on the likely-hood of acceptance”. They could attach probabilities to different factors such as: GMAT score or GMAT waived (if undergrad of same school), amount of acceptance for legacy students, citizenship rates, GPA, ethnicity, program prestige, and degree type. The rates for different schools are public knowledge. For example, may can be found here https://www.prepscholar.com/gre/blog/graduate-school-acceptance-rates/ To train the algorithm, there are many free data sets that can be found in places like https://www.data.gov/education/. I think this would be a very useful application of R for the audience of this post. .
A business question that can be answered with the use of a decision tree is whether someone qualifies for a car rental. The data that can be collected is age, gender, credit score, income, occupation, and criminal history. The advice I would give someone regarding how to select the right predictor variables for a decision tree analysis is to make sure that the chosen variables will give valuable insight to the business question and that the variables have a strong correlation.
The business question to be answered is whether a person will pay for public speaking online course based on any of the following combination of variables such as Age, Gender, Educational level, Income, Job title, and Taking course (willing to pay=1 , not willing to pay =0). A decision tree will provide a simple illustration of branching and slitting of these predictor variables to determine the likelihood of a person who is willing to pay public speaking online course. When choosing the predictor variables, make sure to select variables that are relevant to the analysis.
A business question that can be answered using a decision tree is whether an insurance carrier should write a workers compensation risk for a potential client. The insurer should take into consideration previous losses the client has faced in regards to work place injury, the nature of the business the client is involved in, and safety measures the client takes. It would be beneficial for the insurer to use a large set of data (loss history spanning across multiple years) and use predictor variables that give the tree more accuracy.
A business question that can be answered using a decision tree is whether or not an insurance carrier should write coverage for a workers compensation risk for a potential client. The insurer should look into past injuries in the workplace, the nature of the client’s business (for example: if it is a construction company, it is riskier to write coverage because the likelihood of workplace accidents is larger), and safety regulations the company has in place. It would be beneficial to the insurer if it used a large data set (loss history spanning across multiple years) and predictor variables that increase the accuracy of the tree.
A business question that could be answered with a decision tree is whether or not someone was likely to buy a ticket to a sporting event. Good predictor variables for this would be whether or not they’re sports fans, how many hours of sports content they consume a week, what sports they’re fans of, and if they live in a major city (with a sports team) or not.
A business question that could be answered using a decision tree could be “Is it worth investing in a companies stock?” The data that can be collected and used in the decision tree are variables including, the value of the stock, YTD change, dividends, and quarterly revenue of the company. One piece of advice I would to someone selecting the right predictor variables would be to avoid variables about the company that do not have an effect on stock price. For example stock listings or CEO’s name as these do not have an effect on the value of a stock.
A business question that could be answered through a decision tree would be how likely is it for a student to retake a particular upper level coruse. The data that would be put into this would, the number of students taking the class, their GPA before entering said course, and the grades they receive throughout the course. The right predictor values for answering this question would be the grades students recieved throughout the course.
This is because, if a student with a high GPA, before coming into the course, and is also getting high grades in said course has a higher chance of not retaking the course.
An example could be for a restaurant on whether or not a customer will also purchase a drink once they bought a sandwich. The data needed to perform could be age, sex, location, order price, and order quantity. An analysis like this could be beneficial for marketing purposes for a store to boost sales. Selecting the right predictor variables in this scenario is crucial because one needs to make sure the predictors are relevant to receive more accurate results.
A business that could be answered using a business tree is for when a tour manager of a band is planning a tour. The tree could be used to figure out what cities to play, what venues are in that city, the capacities of those venues and if the demand for the artist is high enough to have a tour date there. This could easily determine for a tour manager what cities have the demand and available venue options for a band to play in that city. The advice I would give someone trying to do this analysis is be as specific as possible with the inputs and make sure to only use the information that applies to the question you are trying to answer.
A business question that could be answered using a decision tree would be: Which customers are likely to purchase concession at a sporting event? The data that would be entered could be age, gender, children and income. My advice for selecting the best variables would be to be sure to include variables that are specific to the question you are trying to answer.
A decision tree can help a retailer determine whether or not a student can register for a class. The data collected for this would consist of the required prerequisites, the number of open seats, the term, sections, credits, and professor. The advice I would give to someone running this analysis would be to ensure that all of the data is correct so that students are not registered for classes they actually are not eligible to be enrolled in.
A business example for using a decision tree could be finding out what demographic is most likely to go on cruises for a travel agency so they no who to market to. They would want to collect data about a sample groups, age, income level, marital status, children, past cruising experience, and how often they vacation. For advice for selecting decision tree variables I would suggest using something that you believe directly will influence the results of the decision. It is better to add too many and have R figure out what is important then have too little and miss a key insight.
A business question that could be answered using a decision tree is whether or not students will take a certain course or not. Several predictors could be based on the financial aid status, they’re need of the course and the time frame of when they could take it
Decision tree could help us determine which kind of restaurant operation will go bankrupt or keep on operating after 1 year. For example, the decision tree could find the location of restaurant who make it and go do not, the kind of food they sale, the operation time ( lunch/dinner), the price range, investment spent on restaurant. All these decisons will make or break a restaurant success.
A business question that could be answered by using a decision tree could be the addition of products to a grocery store. The grocery manager would then factor potential vendors based on their product’s previous success rate and reputation. The decision tree would collect data from vendors like Herr’s for example. This data would include the consumer satisfaction of their products on a scale of 1-10 along with reputation towards other companies they do business with, this would be rated “Poor” , “Neutral” , or “Good”. Another set of data to be included could be costs for transportation and products. All of these variables are important in selecting the right vendor, but the most important variable in my opinion would be product’s success rate with customers.
A business problem a tree could solve is deciding what flavor of ice cream to sell. The decision tree can help determine which ice cream flavor an ice cream shop should have with the tree consisting of demographics such as age, gender, race in the businesses area and base their decision of off their target market in the area and what the target market would buy. Advice is you should be very specific when asking your question that will help solve the problem.