Data Analytics – Section 1

Weekly Question #5

Leave your response as a comment on this post by the beginning of class on April 11, 2016. Remember, it only needs to be three or four sentences. For these weekly questions, I’m mainly interested in your opinions, not so much particular “facts” from the class! If you sign in using your AccessNet ID and password you won’t have to fill in the name, email and captcha fields when you leave your comment.

Answer one of the two questions below (not both):

  1. 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.
  2. What advice would you give someone regarding how to select the right predictor variables for a decision tree analysis?

43 Responses to Weekly Question #5

  • 2) I would advise to choose predictors variables that have the most impact. These include: income, age, has children or not, mortgage and others with the same caliber.

  • A business question regarding your own career that you could answer using a decision tree could be what university you’re going to attend. The data needed to be collected would include if you want to attend college in the first place, where the university is, how much money you’re willing to spend, where the college is, and what major you’re interested in. This would be a somewhat lengthy decision tree needing a lot of information, but it could be done.

  • 1. As a general example, I can imagine a company using a decision tree to answer the question of, “Should we buy or build our technology?” They’d need to collect data such as: unique or common situation, ability to build vs. not able to, cost to maintain vs. cost to buy, etc. This is one small way to demonstrate the many uses of a decision tree.

  • A business question that could be answered by using a decision tree is “Is the job candidate qualified for the available position?” The data need that would need to be collected in this type of decision tree would be whether the candidate attended college. If yes then what degree. Some other data that would need to be collected is whether the candidate has prior experience. If yes then how many years of experience, etc. It would really depend on what type of position it was to determine what data needed to be collected. If it were a tech consultant job, data that could be collected is if the candidate had customer service experience.

  • A business question such as whether or not a company should invest in a new technology can be answered with the help of a decision tree. The data that would need to be collected would be the cost of the technology, benefits, the amount of time it would take to implement, and how much capital does the company have. With the analysis of the collected data, a company should have a general idea of whether or not they should invest in a new technology.

  • Colleges could use decision trees to take a better look at applicants for upcoming school years. The data could include GPA, SAT and ACT scores, gender, ethnicity, family income etc. This could be used to determine who of the applicants to accept and who to reject.

  • What advice would you give someone regarding how to select the right predictor variables for a decision tree analysis?

    I would advise them to break down the problem description clearly, so that they can find their predictor variables easily.

  • A retail company could use decision trees for determining product roll-out strategies. Data which would be included in the tree is previous sales data for similar product categories, geographical and demographic data for the potential new regions in which a product would be rolled out in, and previously existing customer data that the company has, in order to forecast sales predictions in a specific region.

  • 1) 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.

    A business question that could be answered using a decision tree could be used to determine whether or not someone should be accepted to get auto insurance in a certain group. It can be used to determine the age, location, income, and other attributes that can allow the person to be placed into a certain insurance class. If they do not get accepted into a certain class, they will then have to go to the government insurance pool that can be very expensive.

  • A business question that could be answered using a decision tree is whether a company should merge with another company. The aspects that we could analyze using a decision tree would be revenue per quarter, number of employees, number of warehouses, number of products produced, and the cost per product output. Other aspects that could be analyzed that are not financial or number based are the culture of the company and popularity of the brand.

  • One decision that could be made using decision trees is whether or not a restaurant chain should open a in a new location. The manager will be able to analyze the data to determine if the company has the capital available in order to jump start this new restaurant location, and whether or not that particular location would be conducive to generating profit. Data gathered for the decision tree should include company history, customer history, information about the new location (such as traffic patterns, area demographics, crime rates, etc.), and information about the current financial state of the company.

  • 1) A possible real-world problem that could leverage the power of decision trees would be predicting the outcomes of high-risk youth in inner city schools. I know this isn’t necessarily business-related but I think that much of what we learn in this class can be applied to areas outside of business, which is why I like this class. For this problem, you could utilize data on students such as test scores, suspension records, attendance, retention, and other metrics in order to develop a decision tree to predict their likelihood to graduate, their likelihood to get involved with the juvenile justice system, and other possible outcomes for inner city youth. Maybe we already do this, but if we don’t, I think this data would be incredible useful in helping identify students who ned additional supports ahead of time.

  • Business questions: what will be my expected return on investment in the derivatives market in a specific position and economic condition?
    Variables need: volatility, time, interest rate, dividend yield (y/n), market price, strike price, probability of price up-move, probability of price down-move.
    This situation can be examined in multi-period or single-period binomial trees that predict that value of a derivative at a specific time in the future give the above variables. With this information, a trader can then price the derivative accordingly and determine the probability his/her position/portfolio being profitable.

  • A business question that could be answered using a decision tree would be the quantity of products that a store sells. For a store like Dick’s Sporting Goods, who sell a variety of products, they could use a decision tree to break down previous year’s sales and determine which products are most popular based of gender, color, and time of year. For example a pair of Curry or Lebron basketball shoes would probably sell more than a certain toddler shoe since there’s more of a market for Curry’s or Lebron’s so Dick’s would be smart to have more of them on hand than toddler shoes.

  • The best advice I could think of is to first, clearly understand and think it through what you want to analyze. Then write down all of the variables that might relate to what you are trying to analyze. After that you start to choose the most important and most relevant variables and get rid of all others. Variables that don’t have a significant impact on your analysis are probably unnecessary meaning that they shouldn’t be used.

  • When it comes to determining predictor variables, it’s important to carefully examine if causality exists between the variables and the outcome. As shown in class, choosing variables has to be done carefully and slowly; once we rush through and start guessing, we eliminate important variables or include irrelevant ones.

  • After doing research, I realized that lots of financial institutions and companies in various industries use R to carry out functions and bring about various data related to their work and research. Examples include finding standard deviations, averages,maximum & minimum return and even various qualitative information as well. This kind of work will also include decision trees which help in various decision making processes. One example would be, what is the size of investment the company is capable of doing (big or small size)? And according to that, what kind of investments should we look into?

  • A business question that you could answer using a decision tree would be what type of car someone might buy. Car dealers would be able to analyze a person’s age, gender, marital status, and income level to make this prediction through the use of a decision tree. For example, a married man, who has three children, would probably more interested in a safe, child-friendly SUV than a sporty, smaller car.

  • An example of a business question that could be answered using R and decision trees could include whether or not to run a particular ad campaign or not. The data necessary to collect to answer this question could include variables such as the need for an ad campaign or not, cost of the campaign vs. the company’s advertising budget, the time it would take to create the ad vs. when the company needs it to air, and the projected sales impact of running the ad. You could also use a decision tree to filter through which consumers would fit the company’s target market model in order to determine where to focus primary advertising efforts.

  • 1) An auto insurance company could use decision trees to predict how likely it is that someone will get into an accident based off of age, gender, whether they live in a rural or urban environment and much more. This information could be useful to help them determine rates based off of this risk.

  • A great business question that could be answered using decision tree analysis would be insurance companies deciding on who to insure or not and what prices to charge which individuals. Decision trees provide accurate percentages and outcomes which can be a great tool for calculating risk. Also, risk managers could use these percentages to correlate the prices that they should charge their customers based on the amount of risk that is involved in insuring the individuals.

  • A financial team for a firm could use R to better pick an investment location. You could include variables such as NPVs, population/income demographics, and predicted returns to better choose your investment.

  • A great business question would be, how many of our customers in our loyalty program received food assistance from the government? What are the three main products customers purchase the most and what promotional deals can we offer to help support them? The best data that should be collected to perform the analysis is: age, gender, residence, family size, income, and frequently purchased items. To help an employee determine the most important variables, I would encourage them to think about other variables that would influence a customer receiving government assistance: income, family size and gender are all important variables in determining this.

  • Regarding what variables to choose for a decisions tree analysis. Is really to know if those variables would affect the decision tree in any way. Some variables might have no effect on the tree so those variables are useless to a decision tree. The variables that have effect on the decision tree must be added for a more accurate decision making. We always seek having a better analysis an more accurate analysis.

  • 1) A business question that could be used using decision trees is which sales people a company should keep, fire, and promote. Data that should be collected should be sales rate, average sales, average item per order, and other measures of productivity. They can also use office ranking and sales activity throughout the year into account.

  • A business question that we can answer using a decision tree would be if we would want to continue a product line. We can base it off of sales, is it profitable, average order, how many was bought this year, who bought it.

  • A business question for colleges/universities that could be answered using a decision tree could be “Which prospective students should receive a scholarship?”. The data that would be collected to perform the analysis would include (but not limited to): ethnicity, gender, family income, age, declared/undeclared major, High-School GPA, and SAT score. Analyzing this data would regulate whether the candidate gets accepted or not.

  • A business question that we could use decision tree to predict results is does having more advertisement slots on television increase sales? Some data that would need to be collected include duration of advertisement, channels, time of the day being shown, number of ad slots, customer background (age, income), how many calls to secure an order, average time from customers seeing the advertisement to contacting the business to inquire and/or purchase it.

  • A business question for retail stores that could be answered using a decision tree could be what products should we offer to potential/loyal customers. The data variables involved could be among-st the likes of, price of item, style of item, age of customer, color of item, etc. Examining the data could describe who you product attracts as well as piking up the trend the customer is currently appealed by.

  • My business question that can be answered using a decision tree would be whether or not a designer should release a new piece to their collection such as sunglasses, wallets, leather goods etc. They would want to exam the age of potential customers, income of those customers, gender, and loyalty to the designer (if they are a repeat customer or not).

  • What advice would you give someone regarding how to select the right predictor variables for a decision tree analysis?

    When it comes to the predictor variables, you want variables that will accurately help provide outcomes. With that being said, when the decision tree is made it is crucial that the predictor variables play a part in the decision making process to determine the probabilities. A predictor variable should only be chosen if it can aid in determining a probability.

  • An example of a business question that could be answered using R and decision trees could include whether or not for a new computer would be popular. The data might need to perform the analysis would be: size, quality, shape, price, weight, function, memory storage capacity, speed, etc. There are so many good predictors that we can use to determine the outcome, popular or not popular.

  • My advice to those determining which predictor values to use for a decision tree would be to choose the variables that greatly divide the data first. You would then incrementally choose lesser differentiators until you reach the leaf node values. The order of the predictor values would be different depending on what you are trying to decide.

  • Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis?
    My examples is whether or not the company uses the new applications or system for internal management. So it will test the compatibility on the devices (like portable laptops, smartphones, computers and multimedia), price (for installment and permission), privacy and security (safe=additional firewall, special keychain or authorization), complexion (easy or hard to use), range (limitation).

  • 1.) A business decision that could be made using a decision tree is whether to stay in your current role at your place of employment or explore other career opportunities. Data that can be collected for to make this decision is income, daily/weekly hours, culture, benefits, and location.

  • In order to select predictor values in a decision tree it is easiest for me to work backwards. Think about the end goal. What is it that you want to find? From there, you can then backtrack in the data. How will your desired outcome get extracted from the larger data set?

  • An example question I thought of could be do people like to ski or snowboard more and what makes that so? We have learned that some factors that we thought would never play a part in statistical patterns actually do, so there could be some tellers we would never expect for this question. Some could include Income, a comparison of two hobbies picking one or the other (would you rather play basketball or hockey?), what temperature do you keep your house set to? (snowboarders may enjoy colder temperatures than skiers. These could all be right or wrong, but the point is that some indicators could tell a story that we never thought could work.

  • 1. Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis? A business can use decisions tree to see if a person is insurable for health insurance. By looking at different factors you can see if a person is too much of risk. This will help companies choose good risks and not catastrophic risk.

  • 1) A business decision that could be made using a decision tree could be selecting and adjusting types of food to be placed on a restaurant menu. Data that can be collected and utilized for this decision would be highest selling dishes, costs, popularity, taste and ingredients. These factors will help determine which foods will sell better depending on the time of year.

  • You can use a decision tree to make scenario decisions. For example, if you are unsure of how to proceed on a project such as building a new factory or not, using a decision tree would help. You can take old information about other projects and use that to plan for the future project.

  • 1) A business could use decision trees for many things, for example marketing a new drug to a certain geographic location. Whether their return on adverting would be worthwhile to deploy marketing in that area. Data used for this analysis would be, age, gender, income, average usage of drug or similar drugs, etc.

  • Most Hollywood movies are shown in almost every theater around the US. Smaller indie films are only shown in select theaters around the country so a decision tree can be used to decide whether or not a customer will see a certain movie depending on their location, age, sex, neighborhood cluster or other demographics.

  • One business decision you could describe using a business tree would be determining what kind of car you wanted to buy and depending on the price, parts, year, used or new, etc. The key to picking the right variables is to see which one’s would have the greatest impact on the decision. For example price, safety rating, year, condition are all major factors that go into deciding what kind of car you want to get.

Leave a Reply

Your email address will not be published. Required fields are marked *

Recent Comments
Subscribe to Blog via Email

Enter your email address to subscribe to this site and receive notifications of new posts by email.

Join 25 other subscribers