Instructor: David Schuff

Weekly Question #10: Complete by November 30, 2017

Leave your response as a comment on this post by the beginning of class on November 30, 2017. 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.

Thanksgiving is coming!

Name and describe a Thanksgiving-themed question that you could answer using clustering. What data would you collect to perform the analysis? Don’t use an example we’ve covered in class.

57 Responses to Weekly Question #10: Complete by November 30, 2017

  • How big of a turkey (lbs.) should you buy if you invite X amount of people?

    If you know how many people are attending your Thanksgiving dinner, clustering could help determine how big of a turkey you must purchase. The “weight of turkey in lbs.” could be on the x-axis and “# of guests” on the y-axis. From there, you could look at the centroid closest to the amount of guests you invited so that everyone has enough turkey for dinner!

  • What is the amount of turkey consumed based on age?
    Survey data is appropriate for this clustering example. Through surveys taken, age of the person and amount of turkey consumed can create clusters. This data could be used by someone who is preparing to serve a group of people for Thanksgiving, and they could determine how much turkey to buy depending on the age demographic of the guests.

  • Based on age, how much turkey will someone eat on Thanksgiving?

    You can collect data from Thanksgiving dinners this year about the amount of turkey a person eats during their meal. You can categorize the results by age and then use clustering to determine how much turkey certain age groups consume for Thanksgiving.

  • How many dishes should one prepare if one invites a certain amount of people?
    The “number of dishes’ would be on the x axis and the “number of people attending dinner” would be the y-axis. By doing so one can figure out how many dishes to prepare. By looking at the clustered closest to the number of people attending dinner, one can figure out how many dishes to make.

  • A Thanksgiving-themed question that could be answered using clustering is how much food will be left over after the meal. To answer this question, one would need to collect the amount of food prepared in pounds and the number of guests and their sizes. This question would be good for a clustering analysis since people who prepare similar portions compared to the amount of people they had would have a similar amount of leftover food, therefore forming clusters.

  • A Thanksgiving-themed question that could be answered using clustering is how much do people eat based on their weight? If they’re bigger do they eat more, and if they are smaller do they eat less? To answer this, you would need to collect everyone’s weight in pounds and how much food they ate at Thanksgiving in pounds. The X axis would have the weight of the people and the Y axis would have the weight of the food. If people who weight more eat more and people who weigh less eat less, then there will be clusters.

  • How much stuffing might be needed for a big (10+) family?

    To answer this, one might identify the age groups of the family members to identify the portions each person might plan on taking during the meal. By doing so, and clustering them by age, one can identify the amount of stuffing it may take to feed the entire family.

  • One Thanksgiving-themed question that can be answer by using clustering is to determined group of people who really likes to eat Turkey as Thanksgiving dinner, or people who did not eat Turkey at all, or people who eat any meal served or vegetarian. One possible data to collect is create an event page on social website such as Facebook and make a poll for the invitees to vote what meal to serve for the dinner. Then, a cluster of four different group of personal meal preference can be made with clustering.

  • A bakery could figure out how many pies to bake using clustering. The data would need to contain people’s preferred quantities of different types of pie. If they look at the groupings for each type, and their average annual sales, they could discern the amount of each type of pie to bake.

  • What is the ideal time to serve food on Thanksgiving?
    In my opinion, the two most important variables of serving time is the age of the guests – older folks like to eat earlier and guest travel time because if guests have to travel farther they will want to eat earlier. There is an additional wildcard variable, which is the NFL games that day. Watching NFL games is a Thanksgiving tradition, and depending on your family you may want to serve food during the best game of the day or during the worst game so the guests do not miss any of the action. We could make a few different cluster charts to find the ideal serving time.

  • Based on gender and age, how much turkey will an individual consume on Thanksgiving? To perform the analysis, I would collect genders, ages, and turkey consumed in oz. from various individuals. I would then enter the data into R to perform the analysis.

  • A question regarding Thanksgiving that can be answered using clustering could be what time will someone eat Thanksgiving Dinner based on number of guests? The time of the day could be the x-axis and the number of guests would be on the y-axis. The results could be organized by number of guests through clustering, and the information could be used to gain insight on how different size groups eat on Thanksgiving.

  • Question: How many pies were bought based on flavor?
    Data: Pies were sold in total, Pies based on spice were sold.
    From this data, we can know clusters represent which flavor is the most popular one for the Thanksgiving season (pumpkin, apple, etc.)

  • A Thanksgiving question that can be answered using clustering is; how much food will a person consume on Thanksgiving. To answer this question, one will need to gather information such as the average amount of turkey usually consumed and a demographic measure such as age or height/size. This question is good for clustering because clustering allows us to group people by the amount of turkey they eat.

  • A Thanksgiving-themed question that can be answered by clustering would be “How many turkeys of a particular weight will be sold from a particular grocery store?” The data that needs to be gathered include Household Income, and Household Size. The weight of the turkey will be on the x-axis and the number of turkey sold will be on the y-axis.This will allow us to find out turkeys of what weight has the highest demand.

  • A Thanksgiving question that can be asked is, Which s the appropriate sized turkey to purchase for this group of guests ?
    Variables include the amount of guests who will eat turkey, those who don’t eat turkey and prefer a different meat, those who do not eat meat, and the average amount of turkey each turkey-eater would want.

  • How much turkey is needed to serve all invited guests?
    This question can be answered by gathering data on the age, weight, number of guests, and whether a person does eat meat. Clustering will group the individuals based on the amount of turkey they would consume and determine how much turkey in needed to satisfy these guests.

  • Based on the number of children who attend, how much mac and cheese was consumed?

    Data: mac and cheese (lbs), Number of Children ages 0-18

  • One Thanksgiving question that can be asked is which size portions could you serve the turkey. You would need to gather tons of information including how many people are in the setting, the amount of people in the vicinity, the amount that they weigh, how old they are, and the amount of other dishes that there are. It would be most interesting to focus on how much other food is at the dinner on deciding how much turkey to serve per person.

  • Clustering could be used to answer “How much food (turkey + sides) and drinks will I need to feed all of my guests?” Data we would need would be number of people and amount of food. We could also look at age as kids don’t eat as much as adults do.

  • A Thanksgiving question that can be analyzed using clustering is trying to figure out groupings of consumers based on the price of the turkey they bought. To answer this question, that data that needs to be collected includes income, age, family size, and geographic location. This might segment the population to find the customer groups that pay more or less for their Thanksgiving turkey.

  • A Thanksgiving question that can be answered using clustering is how many seats a family might need to have for the infamous “kids table”. In my family, all of our cousins and relatives usually get together for Thanksgiving, and the younger cousins tend to be at a separate table from the older cousins and adults. The data gathered would include how many guests there are going to be for dinner and each guest’s age. By clustering people into age groups, you would be able to figure out how many people total will be attending, so then if not everyone can fit at one table, you could determine how many more seats would be needed for the kids table.

  • One Thanksgiving question that clustering can answer is who ate turkey this year. A lot of millennial decided to have vegan or meatless thanksgiving and age might be able to determine it. The data that can be collected is who ate turkey, age, gender, income, daily calorie intake, and weight (maybe turkey makes you weight more at the end of the day).

  • One question clustering can answer is how much food will the kids eat during the dinner as opposed to the adults so that you can determine how much food to make. Data would be collected on age and amount of food consumed. Clustered segments will determine how much food to make based on the age groups of people attending.

  • One question that could be answered is how much turkey someone eats based on their income. The data that would be collected is turkey consumption in lbs and annual income. Clustering would help visualize what income levels are more likely to buy more turkey.

  • A Thanksgiving-themed question that could be answered using clustering is what is the size of turkey need to prepared for inviting x amount of people to the dinner. To perform the analysis, data such as numbers of people that will eat turkey, people that don’t eat turkey, amount of vegetarians among the group, and the average amount of turkey will be consumed by each turkey-eater.

  • A Thanksgiving question that can be answered using clustering is how muchTurkey will males versus females eat on Thanksgiving? This would give people an estimate of how big of a turkey to buy. I’ve always heard you should buy a turkey big enough to feed a pound a person, but is that too much? To answer this question, the data collected should be the how much turkey in pounds each male and female ate at the Thanksgiving.

  • A thanksgiving question that clustering can answer is which food is more popular/demanded than others. Data such as ages and amount of food consumed would be collected, and based one the clustered segments you can see what food is more popular and by what age group.

  • A Thanksgiving-themed question that could be answered using clustering is how much to make of each side based on the popularity of each side. You could collect data on the amount of people who will be attending Thanksgiving and which sides are known to go the quickest to perform your analysis.

  • How much turkey (lbs.) should one purchase based on how many guests are attending?

    Data collection would be based on number of guests and size of the turkey. Clustering the data would determine how much turkey should be purchased if n guests were attending.

  • How long does a person nap for after eating their Thanksgiving meal? Data collected would be the amount of food consumed (lbs) and the time spent napping.

  • For a bakery, one Thanksgiving question that clustering could answer is “how many pies to make based on flavor?”. The data would depend on customer’s choices of pies flavors and their quantities. Then, based on the collected data, the quantity of each flavor is determined.

  • One question clustering can answer is who eats cranberry sauce by age. Many younger people do not like cranberry sauce and many older people do enjoy it, so clustering the data of how much cranberry sauce was eaten by the age the of the person eating could group people and see if there is a trend to higher consumption based on the age of the people at thanksgiving.

  • Something that you could use clustering for in relation to Thanksgiving would be to estimate how much turkey an individual will eat based on their weight. You would need to collect data on people weight and the amount of turkey that they consume over previous Thanksgivings. Then you can cluster into general weight classes (ex. 130lbs- 145lbs, 146-160 lbs, etc). We can then examine this data and approximate how much turkey you would prepare based individuals weight and your predicted consumption.

  • A Thanksgiving-themed cluster could be how many servings of mashed potatoes each person eats based on their height. The data collected would be each person’s height and the number of mashed potato servings they ate.

  • How many political arguments get started per political affiliation (Democrat, Republican, Green, Libertarian, Centrist) at the thanksgiving dinner table.

  • A thanksgiving-related question that could be answered using clustering would be what time will guests go black friday shopping relative to the time their thanksgiving dinner ended? Some of the variables that might be examined include the time of the meal, portions the guests ate, and how long they spent with friends and family before going out to shop. This could also be looked at over the span of several years to see if there has been any difference in recent years in the amount of time consumers are spending with loved ones before they go out to shop.

  • Question to ask would be how much leftover food should I have at the end of Thanksgiving so the fam doesn’t complain that I sent them with none. data collected would be the number of guests, and the amount each of them ate at past Thanksgiving dinners then cluster in the amounts. After that, just cook more and send everyone off happy.

  • A Thanksgiving-themed question that could be answered using clustering would be something like, ‘how much pumpkin pie is eaten based on age, gender, ethnicity, etc.” I would say you could measure the number of pies eaten as slices. As for the amount of clusters I would use, I would say something like 3-5 because I can’t imagine there being a lot of variance, seeing as how everyone loves pumpkin pie!

  • You could use clustering to figure out how early you’d need to leave on Thanksgiving to get to the stores early enough so that you don’t need to fist fight someone for a PlayStation (although it’s still an option for the more adventurous types). First you’d want to collect data on opening times for stores in the area, and then you’d want to look into how early lines have started on previous Thanksgiving days. After collecting all of the data, you could find out how early people start showing up generally so you could be the first one through the doors when the store opens.

  • A Thanksgiving question that could be answered using clustering is how many pies of each flavor are bought at different stores. The flavors could include pumpkin, apple, pecan, cherry, etc, and the stores could be different bakeries or different grocery stores.

  • You could ask how much pounds of food would be consumed per person during the Thanksgiving meal. The data collected would be weight before and after the meal is served.

  • My thanksgiving-themed question concerns finding mutual preferences in food (turkey, mashed potatoes, green beans, stuffing, gravy) amongst guests. My collection of data would hold the number of guests eating, and the weight of each respective portion consumed by that particular guest. Clusters are aggregated when similar portions of similar choices are consumed by guests.

  • Question: How much stuffing should be made?

    Data collected: How many people will be attending, how much stuffing was made/eaten in previous years

  • A Thanksgiving cluster could be created in order to show, per household, the number of pounds of turkey purchased and the number of guests invited for dinner. Some people aim for a lot of leftovers, some try and guess the right amount, and some may run out of turkey before everyone has had some. Many hosts are unsure of how much food to buy and prepare and the turkey/guest ratio cluster could give insight into how much to buy for next year.

  • A Thanksgiving themed question that we could answer using cluster would be “What is the weight of a turkey needed to be purchased and cooked to serve a certain amount of people?”. The data collected to answer this question would be the number of guests that will consume the turkey, how many people like white meat, dark meat, and how much they have eaten in past dinners.

  • A thanksgiving themed question would be how much turkey someone would eat based on age. The data I would collect is the ages of the individuals, the weights of the individuals, and the amount of turkey consumed by the various individuals. I would also compare it with past data to see if the outcome would change based on the year we are in.

  • A thanksgiving themed question could be, “How much turkey do people eat and what is their corresponding weight”. So this could be for tons of clusters or even select few. The variable would be turkey and weight. You could do 100 pounds, 150 pounds, 200 pounds, 250 pounds, and 300 pounds. This would be a select area, but you will be able to see how much food people consume and what their corresponding weight is.

  • A large turkey distributor that services a group of small towns might want to know “How many deliveries of turkey will be needed by each serviced grocery store?
    Some data they would need to collect could include: town population, number of “family” households, grocery stores per town, grocery store monthly shopper count, etc.

  • A Thanksgiving-themed question that you could answer using clustering is how many days will it take a family to finish the leftover from thanksgiving. The data I would collect to perform the analysis are the the amount of people in the family, the amount of leftover they have, and the amount of dinner food they normally eat per day.

  • A Thanksgiving themed question that one could ask is how much turkey will a particular family eat. You would want to make clusters based on how many people will be there, their ages, and how much other food there will be. You could also take into consideration the numbers from the past few years of Thanksgiving dinner.

  • A thanksgiving based question that clustering would answer is how much pie is going to be consumed?
    To answer this question you will have to gather data on how many people are coming and data on previous years on how much pie was baked and consumed.

  • A thanksgiving based question that you can use clustering to answer is how heavy is someones thanksgiving plate in comparison to their weight. To answer this question, you can gather data that includes age, weight of person, types of food, weight of food, how much was made/eaten per person.

  • A thanksgiving based question that you can use clustering is how much weight a person’s thanksgiving plate is in comparison to their body weight. To answer this question, you can gather data including age of person, weight of person, type of food, weight of food, and amount of food made and consumed.

  • One question that could be answered is what type of dishes are prepared for thanksgiving based on location. For example, some people may prepare things other than mac and cheese, cranberries, etc. The data to collect would be the types of dishes, and how many times they each occur at each location geographically.

  • A thanksgiving themed question could be the size of the turkey in pounds based on the amount of guests at the dinner. Some variables you could look at for this could be the amount of guests, their ages, how many are vegetarian or prefer a different meat and the weight of the turkeys.

  • How many pies might be needed for a family of 18?
    Consider the age and the average portion size each person eats. cluster this data by age so there is enough pie for all!

Leave a Reply to Tonatiuh Paredes Cancel reply

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

Where and when do we meet?
Alter Hall 232
12:30 - 1:50 Tuesdays and Thursdays
Office Hours
David Schuff (instructor):
2:00-3:00, Tuesdays and Thursdays
Speakman Hall 210E and email (see my site)

Lauren Soentgen (ITA):
1:00 - 2:00 Mondays
11:00-12:00 Fridays
Speakman Hall 210D (and see her site)