Reading Assignment #2 – Due 1/18/12

Here’s your next reading assignment. Read Section 1.4 in your textbook and answer the following questions by 8 a.m., Wednesday, January 18th. Be sure to login to the blog before leaving your answers in the comment section below.

  1. Why do you think that there’s only one pie chart in the textbook?
  2. What components of each plot in Figure 1.35 do you find most useful?
  3. What’s one question you have about the reading?

70 thoughts on “Reading Assignment #2 – Due 1/18/12

  1. 1. Pie charts are more difficult to analyze. Unless labeled, they provide you with no sense of the numbers behind the graph (as an axis might). They also make it difficult to compare between groups. While its easy to see that groups Small and Midsized are close, its very uncertain how much greater they are than Large.

    2. I think the box and whisker plot is valuable for comparing the mean and extremes of the three categories. However the hollow histogram is great for showing the market distribution (and looks really cool).

    3. Would the author (or you) agree pie charts are useful for graphing a budget? I think it can be a very literal interpretation with valuable information if labeled correctly.

  2. 1. Pie charts are not especially helpful from a statistical point of view. Organizing the data in that format makes it much more difficult to draw information about the group sizes, and without horizontal and vertical axis it is much more difficult to integrate.

    2. the side by side box plot is very useful for visualizing trends in the data whereas the hollow histograms give a much better idea of the distribution.

    3. I thought this reading assignment was very straight forward.

  3. 1. It is because pie charts are not as useful as other types of representation as it is difficult to recognize the actual level of each portions.

    2. Box plot: The outlier, the location of the median line relative to the size of the box and the whiskers.
    Hollow histogram: All three price distributions for each type are being stacked together and makes it easy to compare them to one another.

    3. In what situation does a mosaic plot become more useful than the segmented bar plots?

  4. 1) Pie charts are not as useful as other charts, since it is more difficult to compare group sizes in a pie chart than in a bar plot, for example.
    2) In Figure 1.35a, it is easy to see the variability in the data of the price and car size. For example, the price of midsize cars seem to vary more than the price of small cars. Furthermore, the interquartile range and median are defined, and outliers are clearly marked with a (gray) dot. In Figure 1.35b, it is clear to see the variability in the price of the car sizes and the modes are easy to discern as well. Since the x-axis is the same for all 3 car sizes, comparison of the prices across the sizes is somewhat easier than in the side-by-side box plot.
    3) Is the y-axis on a hollow histogram, the frequency that the observation is seen (counts) or is the proportion of the observation? (In any case, this yields the same heights, but I was curious as to the actual label for the y-axis)

  5. 1) There is only one pie chart in this book because as the book states, pie charts are not as useful as other charts in a data analysis; so the book just shows this one as an example but won’t use pie charts any further.
    2) In the side-by-side box plot I find it useful to be able to see the median of each data set, the ranges of data, and the potential outliers. In the hollow histograms I find it useful to be able to see the mean of data sets and the skewness.
    3) I have never seeing mosaic plots before. Are they less useful and when would it be good to use them?

  6. 1. The function of a pie chart is to show the percentages of values for different categorical variables. However, a proportional bar plot or mosaic plot can represent the same data more intuitively and more accurately, rendering a pie chart generally ineffective and less useful than other forms of data representation.

    2. The side-by-side bar plots are useful for showing where the medians of different data sets lie in relation to one another, while the outline histograms give a good intuitive picture of data density – where each data set peaks compared to the others.

    3. Is cluster sampling effective for garnering a random sample?

  7. 1) Pie chart is not very useful at data analysis compare to other techniques. It is kind of hard for readers to get information that we want to deliver.
    2) I think the midsize because it has a wider range of price value compare to the other two
    3) Question that I has is how do we read information out of the side by side plot or the hollow histogram. It is kind of confusing to me.

  8. 1) There is only one pie chart in the textbook because it is difficult to glean information from them, as it is relatively difficult to compare the sizes of different sectors.

    2) Both the side-by-side box plot and the hollow histograms seem kind of busy; however, both display some clear, useful data. It is easy to see the median price for each type of car on the side-by-side box plot. On both charts it is easy to see the range of prices for each type of car.

    3) 3D data visualization? This is not so much a question about the reading, but what future do you think 3D data visualization has? Eyeing the future, one could see where 3D visualizations would be visually frustrating and contain too much information to digest. On the other hand, if done well, they could be very useful and visually pleasing.

  9. 1) Because it is hard to tell which of two similarly-sized segments is actually larger.

    2) The box plot is useful in seeing where the median occurs in the distribution and whether it is fairly centered or if it is skewed. The hollow histogram is useful in showing the frequency of the cost ranges across the groups.

    3) None, I found the reading to be fairly clear.

  10. 1. Pie charts can be easily misinterpreted, due to the nature of how they are drawn. Other types of charts can display the same information with much better clarity.
    2. For the first plot, I find the median bar on the plots to be the most useful. On the second plot, the key for the graph is the most useful.
    3. Why do we have so many types of graphs? Is there a most useful graph?

  11. Q: Why do you think that there’s only one pie chart in the textbook?
    A: They aren’t as useful as other charts in a data analysis because it is difficult to compare group sizes in a pie chart relative to other types of charts.

    Q: What components of each plot in Figure 1.35 do you find most useful?
    A: In the box plot, the median, IQR and whisker components are the most useful because one can easily look at the plot and tell what’s going on. The hollow histogram makes it easy to compare small, midsize and large cars because the histogram overlap and allow you to visualize the data better.

    Q: What’s one question you have about the reading?
    A: Row proportions vs. column proportions when dealing with contingency tables? What situations would you use each?

  12. I think there is only one pie chart because they can only display one category per chart and it’s hard to compare sample sizes. I don’t understand what the charts in figure 1.35 are showing.

  13. 1. Because it’s difficult to tell the proportions of different categories while looking at a pie chart, while if you use a bar graph instead, you can easily tell the comparative size of each category.

    2. The side by side box plot is useful for seeing the IQR and outliers for each type. The hollow histogram is useful for seeing how much each size of car overlaps in certain price ranges.

    3. Are there any situations in which pie charts are better than bar graphs?

  14. 1. The main problem with pie charts is that the data is really hard to compare. Sure you can easily see dramatic differences, but when two sections are close it’s almost impossible to tell.

    2. The box plot has elements in it which I found the most useful. Looking at it I instantly have a better idea of how much an average car of each type might be, and how flexible that price point is. It doesn’t take much thought to get its point across.

    3. Are pie charts really considered that bad? I mean if a heavy weight towards one category is being conveyed to prove a point then I’d think that it’s a very useful tool. Everyone gets them instantly so that helps it commercially, which is a plus. Sometimes the slight differences don’t matter, just a general idea needs to be displayed, and these seem good for that.

  15. Why do you think that there’s only one pie chart in the textbook?
    > Because bar plots give the same information in an easier to read maner.

    What components of each plot in Figure 1.35 do you find most useful?
    > I think the side by side box plot is the most useful.

    What’s one question you have about the reading?
    > I don’t totally understand the mosaic plots, especially figure 1.32.

  16. 1. Pie charts don’t really tell us precise proportions and make it difficult to compare the sizes of different pieces of the pie. It would also be really hard to display trends, correlation, or changes over time using just a pie chart, as graphically pleasing as they might be.

    2. The side-by-side box plot makes seeing the distribution of the price of each type of car easy. For example, the box plot for midsize cars is far wider than the ones for small or large cars and has a clear outlier, information that is not as easily visible in the hollow histogram mostly because the dashed lines are difficult to distinguish at first glance.

    The hollow histogram makes it easier to tell where the mode price for each type of car is located and also makes visualizing the shape of the distribution far easier.

    3. When are mosaic plots really useful? They just seem difficult to understand to me.

  17. 1. A bar plot is better in visualizing group sizes than pie charts.

    2. the box plots and the hollowed histograms are useful in representing the variability of the data.

  18. 1. There is only one pie chart in the textbook because the chart type fails to convey information effectively. For example, in Figure 1.33 we can see that midsize and small cars have approximately the same amount in the pie chart, but we can tell for certain that they are different amounts when we look at the bar graph. A pie chart is an ineffective way of portraying information and the textbook would rather not use it to convey information in the hopes that the student will not use it as often.
    2. The side-by-side box plot in Figure 1.35 allows us to see the exact median price of each group of cars, it also allows us to see the relative distribution of each car type. For example, we can see that small cars do not vary much in price whereas midsize cars vary greatly. Also the large group of cars below the median price varies very little, but above the median price varies greatly.
    The hollow histogram in Figure 1.35 we can tell the relative amount within each price range of each car type. From the graph, we can tell that the graphs are unimodal and the amount in variation of price. The graph suggests that midsize cars right skewed and large is also right skewed, though to a lower extent. The small size cars are uniform in their distribution. We can also tell that midsize cars vary greatly in price while small cars vary the least with large cars in the middle.
    3. Mosaic plots are useful for showing information such as amounts of a certain type relative to the sample size, but I do not see them around that often. Is it because it requires prior knowledge of how to read it in order to comprehend it? Aren’t graphs that do not require prior knowledge better for distributing data to more people?

  19. 1) Pie charts are visually pleasing but poorly convey data, since proportions are difficult to judge
    2) Plot 1 shows outliers well. Plot 2 is simply confusing, with overlapping data and an undefined y-axis, thus I have nothing positive to say about it…
    3) Aren’t there at least SOME cases in which pie charts are useful? Like when visual appeal supersedes practicality in a particular data set?

  20. 1. Pie charts are not as useful as other charts and diagrams are for visualizing data. The example in the book shows this clearly: it is much easier to see a small difference between two data types using a bar graph or histogram than it is to see that difference using a pie chart.

    2. I like that the hollow histogram shows all of the histograms for the three categories in one plot. The overlap in each gives a way to quickly see how similar the distributions in price are for each of the categories. The outliers shown by the side by side box plot are helpful as we may not want to include such data in our calculations (since extreme circumstances may have caused these data points to be so far separated from the rest of the data). It is also helpful to be able to see where each of the medians (and first and third quartiles) lie.

    3. How often are mosaic plots used? The segmented bar plots seemed much more immediately accessible than did the mosaic plots. I’m wondering what data might be better visualized using a mosaic plot over a segmented bar plot…?

  21. 1. There are likely a couple reasons why the authors do not support pie charts.
    Reason 1: Pie charts do not provide exact amounts and thus can’t be used for data abstraction, only for proportional visualization of a single data type.
    Reason 2: Pie charts are limited to providing only one data type of information as a proportion, unlike the graph shown in the video in class which had 5 data types!
    Reason 3: Pie charts can be difficult to read if the pie portions appear similar size… a narrow bar chart or line graph provides a much clearer visualization as to how the two portions compare.

    2. Useful components of the Side-by-Side Box Plot:
    a. Outlier visualization helps to clarify which of the cars are “Bentley’s, Alpha Romero, etc.” If you were someone looking at cars for the general public, this information would be helpful since it would weed out those car choices.
    b. The IQR is very helpful to give a sort of “average” sense or middle 50% suggestion as to how much each type of car costs. This is very useful when deciding which car to go for.
    c. Only the same lines the range markers are very useful in showing how much you can pay for a car of a certain type, or how little.

    Useful components of the hollow histograms:
    a. Its nice to see the entire distribution, skewness, nodes, and overall size.
    b. The bins provide a great deal of information regarding all the possible prices available for a certain type of car. For example, the small car has very narrow range and only covers a couple bins, so we know that if we buy a small car, it can only be within a limited price range, and if that price range doesn’t suit us, we shouldn’t buy that car. This is very helpful.
    c. Laying the distributions over top of each other allows for easy comparison.

    3. I would love more practice on questions like these. Researchers suspect the proportion of male possums might change
    by location. A contingency table for the pop (living location) and sex variables from
    the possum data set are shown in Table 1.28. Based on these researchers’ interests,
    which would be more appropriate: row or column proportions?

    I always find the wording misleading and pick the wrong answer. How can one getting better at this?

    Also, I found it highly ironic that the authors don’t like pie charts, but they like mosaics (which in some sense, are really just two dimensional pie charts). Mosiacs are definitely better than pie charts, but they have a lot of the same limitations, so why would they be worth mentioning?

  22. 1. There is only one pie chart because they are not as clear as histograms and bar graphs. Although the same data may be available, pie charts make it more difficult to visually see the difference in data that has little variance.

    2. Both charts show variance pretty well. I think the histogram makes the mode for each type of vehicle more apparent. The box plot shows the average pretty well.

    3. I understood this topic pretty well and don’t have a question.

  23. 1. Pie charts are all around poor visual models. First, they take up a lot of space (bad real-estate) and give minimal information. Second, the information they give is all relative, as in it can only compare discrete variables. Thirdly, those discrete variables have to be relate to ‘parts of a whole’, which not all data can do.
    2. For ‘type’, I like how simple and easy the graph is to read instantly; there are no unnecessary colors or lines. The ‘price’ gives a better feel for how the data is spread out, as my brain can more easily understand where the mean is and how the data is spread. Unfortunately for pice, the dotted lines get a little complicated and could confuse colorblind users, or even non-colorblind users.
    What’s one question you have about the reading?
    3. Which graph is the easiest for users to determine relational data, numeric data, and other types? What does each type of graph specialize in for user-readability?

  24. 1) Pie chart is harder to analyze when two or more similarly sized groups of data are present.

    2) For the left chart – The size of the box, which can represent variance and/or standard deviation.
    For the right chart – The height of the histogram, which can represent mode.

    3)How is a mosaic plot used in real practice? What are some examples?

  25. 1. Pie charts are not really useful for data analysis. In the example, it is hard for us to determine which car size has the largest count as the midsize and small portions look similarly big from the pie chart. From the bar plot, we see that midsize type cars has a slightly more count clearly.
    2. I find the box plot to be effective in displaying the median and the spread of the data. The histogram is excellent in displaying the mode.
    3. I am still not clear about the mosaic plots. Other than finding the median and the spread, is there other ways to analyze the data through box plots?

  26. 1.
    It is more difficult to compare group sizes in a pie chart than in a bar plot, so they are typically not as useful.
    2.
    For the box plot, it is most useful to see the small, midsize and large price ranges clearly apart from one another, while being able to see the distribution for each one to compare to one another. For the hollow histogram, it is useful to see how the 3 car types prices overlap one another.
    3.
    Are there any reasons you would use a mosaic over a segmented bar plot (or any plot for that matter)? The mosaic seems like a less clear and confusing representation of data, and I can’t really see any reason for representing your data in this way over other types.

  27. Why do you think that there’s only one pie chart in the textbook?
    a pie chart is not very useful since it doesn’t make comparing group sizes as easy as other charts such as bar graphs

    What components of each plot in Figure 1.35 do you find most useful?
    I think the hollow histogram is more helpful because it is more intuitive to me, and it also shows the price jump in large cars from 40 to 60 with almost no 50 which is not possible in the side-by-side box plot

    What’s one question you have about the reading?
    is the front and rear proportions in figure 1.32 independent from the size of the car?

  28. A)
    In the left plot (Side-by-side box plot), the most useful compent is “how the plot shows the mean and medain of data values in a box”. Also, I like how it shows the highest and lowest values as limits.
    In the right plot (Hollow histograms), the most useful compent is the information zone.
    B)
    Because pie charts are not typically as useful as other charts in a data analysis. Also, it is more difficult to compare group sizes in a pie chart than in a bar plot.
    C)
    In general, the topic is clear 🙂

    Thank you

  29. 1. Pie charts make data comparison much more difficult than is necessary. Many details of the data are left out of a pie chart which other graphs are able to successfully show.
    2. The side-by-side plot allows us to easily determine where the majority of the prices are, and to quickly identify outliers. The hollow histogram allows us to effectively see how the data overlaps and to see similarities between the data sets.
    3. What is the point in using a segmented bar plot over a mosaic plot?

  30. 1) Pie charts are not a good way of visualizing data, because they make it difficult to compare group sizes to each other.

    2) The box plot makes it easy to visualize the mean, range, and quartiles of the data, as well as any outliers. The histogram allows to visualize the general distribution of the data, and how it compares to the rest of the data.

    3) Why don’t the graphs in the reading have titles?

  31. 1. As stated in the book, pie charts are not the best way to analyze data. It is difficult to compare group sizes on a pie chart. Bar plots are more commonly used and easier to analyze. For this reason, pie charts are not the most effective way to display data.
    2. In the side-by-side box plot, the labeled axes make it easy to see the price difference amongst each group. This chart is easy to read. The hollow histogram is a bit more complex. The hollow bars overlap each other, making it difficult to note the difference amongst each group represented. Despite a lack of clarity, this plot allows a clear distribution of price where each group falls. This makes the data show in more detail.
    3. Why would a hollow histogram be used over a bar graph? Hollow histograms seem to confuse me when I look at it.

  32. 1. Pie charts are not useful because it is harder to compare group sizes than it is with a bar graph.
    2. For the plot on the right, the most useful component is the fact that it shows where the prominent peaks, or modes, are if they exist. For the plot on the left, the most useful component is that the median is shown.
    3. Which of these charts will we be using most in this class?

  33. 1. Why do you think that there’s only one pie chart in the textbook?
    Pie charts are an overused and really aren’t that useful. The same data is easily captured in other visualization displays that show other information as well.

    2. What components of each plot in Figure 1.35 do you find most useful?
    The whisker graph more easily shows a comparison of the ranges of the price between the different types. The histograms are better at showing the relative magnitudes.

    3. What’s one question you have about the reading?
    Why exactly are pie charts so prevalent in spite of their relative uselessness? Is it from early versions of MS Office, Lotus, etc making it easy to make the charts?

  34. 1. Because it is often difficult to eyeball and compare categories that are close in percentage

    2. The side by side box plot really shows a lot of information that can be seen as useful. I like how you can directly compare the distribution and median of each of the different types shown. The hollow histogram again clearly shows the distribution of prices for easy comparison.

    3. Are multiple level pie charts also taboo in the statistics world?

  35. 1. Pie charts are not as useful as other charts in data analysis. It is more difficult to compare group sizes in a pie chart than in a bar plot.
    2. For the left plot, the price distribution of each type is the most useful. For the right plot, the type distribution of specific price level is the most useful.
    3. I think I still don’t quite understand the side-by-side box plot. Can you explain it in class tomorrow?

  36. 1. Pie charts are not as useful as other charts in data analysis.
    2. The mean value of the side-by-side plot is useful as it not only shows the average but also gives a rough trend of the data varies.
    3. No questions

  37. 1. There is only one pie chart in the book because pie charts are not very useful for visualizing data with more than one grouping. A pie chart just shows relative proportions in one group set and does not give accurate representation of all data that may be useful.
    2. I find the way each is scaled to be useful. The histogram seems like a more useful tool in that its easy to see the groups based on size and the amount of cars of the size that were priced at the same value. Both plots are very useful for noticing trends in price as it correlates to increased size of vehicle.
    3. I was a little confused on how exactly to interpret the side-by-side box blot.

  38. 1. Pie charts have no inherent scale while a bar graph has a scale and displays the data in a way it can be compared quantitatively.
    2. In the left panel, the box is the most useful since it provides the two quartiles of data that are closest together.
    In the right panel, the fact that all three histograms are present so that the data can be compared.
    3. Are mosaic and pie graphs the better chart type for those who are uninterested or unable to read and understand the data?

  39. 1. There’s only one pie chart because while it does give a complete representation of the data, the exact values of each section are difficult to determine and compare with each other.

    2. The box plot gives a better view of the average price of the cars and their variance. The hollow histogram gives a better representation of the mode for each car size.

    3. Is there a way to determine the optimal method to display data, or is it just whichever best represents the data?

  40. 1. Pie charts are relatively useless when compared to the all the other methods of data visualization. It is very difficult to judge the relative size of the pie slices. Also, pie charts take up a lot of space to present only a little bit of information. The data in a pie chart can much more clearly be understood in a bar graph.
    2. For the side-by-side box plot I find the median lines as well as the upper and lower whisker lines most useful. All of these points give a complete representation of the trend obtained from the data. For the hollow histograms, I find that the “bars” help visualize and emphasize the density of the data, which is very useful.
    3. What is the point of mosaic plots when you have segmented bar plots available? What type of information do mosaic plots give you that bar plots cannot? I am confused of when I could pick a mosaic plot over a bar graph.

  41. Pie charts can only represent one thing: proportionality. They don’t show large amounts of data, and it’s not organized well. A pie chart, while being an attractive graph, typically needs to be accompanied by other graphs to be relevant in any way.

    For the two figures in Figure 1.35, I find the scaling the most helpful for both. Putting the data all in on the same scale makes it a LOT easier to compare. With the side-by-side box plot, it was nice to see the interquartile ranges comparing costs, and for the hollow histograms, it was nice to see the more specific price points, and how tight or wide the data actually was.

    When would a mosaic plot ever actually be useful?

  42. 1. It is difficult to know the exact relationships between group sizes in a pie chart. A bar graph shows these differences much better than a pie chart because it is hard to look at a pie chart and know exactly how much bigger group A is from the other groups of data on the chart.
    2. I find the size of the boxes in the two graphs to be most useful because it tells where the majority of the prices for each group of cars lies.
    3. What does the black line represent in the side-by-side box plot?

  43. 1. The book only has one pie chart because they are not typically as useful as other charts in a data analysis. It is more difficult to compare group sizes in a pie chart than a bar plot.

    2. In the first plot I find that the most useful component is the midpoint line. In the second plot I find the ranges in prices between the 3 types of vehicles most useful to understand the data.

    3. I have a question about the side-by-side box plots. This is my first time seeing them and I need help trying to read what the graph is displaying.

  44. 1) There is only one pie chart in the book because when comparing data a pie chart is a lot harder to read. That is because a pie chart has the data arranged as a section of a circle, so it’s harder to tell which section is bigger than another section and by how much.
    2) In the plot on the left I think that the fact that the three different type of car is on the same graph. So they have to same scale, which makes it easier to compare them. On the graph to the right I think that the three different lines, representing different types, are really useful. Although it is harder to compare the two dashed lines.
    3) If pie charts aren’t useful then why are they used so much?

  45. 1) Why do you think that there’s only one pie chart in the textbook?

    The pie chart is very inefficient for data visualization since it only displays a single variable and category. Also, for larger data sets, the “slices” are far too small for people to really see. This makes it undesirable with no real upside compared to bar graphs.

    What components of each plot in Figure 1.35 do you find most useful?

    Being able to see where the median is gives a lot of information about the distribution in the box plots. This lets us know that large cars have a very low median despite having some very high priced cars that aren’t outliers. From the box plot you can also tell that the vehicle prices tend to be very concentrated in lower price ranges and less so by high price ranges from the distance between the quartiles and the edges of the plot. The histograms are useful because they give a better picture of the actual distribution, especially within the inter-quartile range.

    What’s one question you have about the reading?

    How do you calculate whether a point is an outlier?

  46. 1. The author outright states that they are harder to read than bar graphs. A slice that differs by only a few percent from another slice will appear to be the same size. With bar graphs, it is generally rather easy to see which is taller.

    2. The hollow histogram is much easier to read in my opinion, because you need less training to read it, it explains itself quite well, while the side-by-side boxes graph requires that you understand the graphs conventions to get all the data from it.

    3. When would a mosaic type graph be useful? They seem more cumbersome than bar charts.

  47. 1) There’s only one pie chart because pie charts aren’t as useful as other graphs. They are great at giving a general outline of group size, and it’s easy to tell which group is bigger but determining actual proportion size is much more difficult than on a bar graph.

    2) 1.35 (a) shows IQR, median, range, outliers, and gives a general sense of the distribution of the data. 1.35 (b) is useful because it displays all three data sets on one graph but it is clear where the sets overlap unlike a solid histogram.

    3) Figure 1.32 took a long time to understand. I eventually understood what they were trying to do and what each bar represented, but I had to keep looking back at fig 1.31 to make sure I wasn’t confused about what each bar represented.

  48. 1. It is more difficult to compare group sizes in a pie chart than in a bar graph, also it gives little information other than percentages relative to a total amount.
    2. the box and whisker plot of each size shows where 50% of the price range lies with and any outliers. Also, the three histograms also show price range with the most frequent prices of each range, giving much more information than the pie chart.
    3. When is a bar chart vs a histogram used?

  49. 1. data can be shown more clearly and usefully in other formats such as bar graphs

    2. I find the side-by-side comparison on the box plot most useful because it provides a clear and direct comparison among the small, midsize, and large prices. I find the hollow histogram most useful because of the comparison of the height of the lines of each type.

    3. The hollow histogram, similarly to many in-class examples we have used, has no “y variable” — how does this affect the visualization of the data — is it better to present the data with or without one? or is it situational?

  50. 1. Because they are bad at data visualization, it is difficult to tell the difference between two categories with very similar data values.

    2. The side-by-side box chart makes it very easy to directly compare the mean of each group. The hollow histogram makes it easy to compare how each category is distributed.

  51. 1) Pie charts are more difficult to read and less precise. Group sizes are harder to detect in pie charts because “pie slices” of similar size appear to be exactly the same size making trends hard to detect. Bar plots distinguish group sizes better.

    2) The first plot easily shows the trend between vehicle price and type. Prices are split into groups which can show one if larger or smaller cars cost more or less. The second plot displays the specific trends of each vehicle type since lines are drawn through each sets of data.

    3) Is it unprofessional to use a pie chart when making a presentation in business?

  52. 1.With the help of bar plot, numerical result can be easily found. What’s more, the proportion of each kinds of car be observed directly from the pie chart. Therefore, one pie chart is enough.
    2, In the left plot, I can find the highest and lowest price according of each kind of car. What’s more, the most important information is that the left plot shows the average price of each kind.
    In the right, I can also find the price distribution of each kind of car. The advantage of right plot comparing to the left is that the distribution of price is much easier to observe.
    3. In 1.35 left plot, I thought that the box mean the major part of price. But how can I determine where the “box” should be drawn? Any standard?

  53. 1. The author of the book contends that it is more difficult to compare group sizes in a pie chart than in various other types of charts Another reason that a lone pie chart appears in this book might simply be because they are overused and tend to be dull. It is clear that there are much more comprehensive, useful, and visually pleasing methods of data representation.

    2. In the plot on the left, I think the most useful piece of information is how the spread of each data set is given. You can clearly see that small cars have the smallest range of prices, midsize cars have the largest range of prices, and large cars have a range in between these two. In the plot on the right, I find the heights of the histograms to be particularlyuseful since it allows you to compare the modes and distribution shapes of the different groups easily.

    3. What is the advantage of a hollow histogram to a regular histogram? How do they differ?

  54. 1. There is only one pie chart in the textbook because as a pose to the other charts for data analysis they are hard to compare group sizes. A bar chart is much easier to read and compare group sizes than for a pie chart.
    2. For the side-by-side box plot, the line that represented the median price for that type car was the most useful. It clearly states the most likely price of a certain type car. For the hollow histograms, the most useful component is the prominent peak of each histogram which shows the most likely price.
    3. Reading the side-by-side box chart seems complicated. The median is very clear but the markings are confusing in what they represent and how.

  55. 1. It is less useful in data analysis than other types of charts. An example of one of its weaknesses is the relative difficulty to see quantitative differences between large sections of the pie chart as compared to other charts such as the bar plot.

    2. The side-by-side box plot is useful in that it shows the median whereas the hollow histograms plot does not. The side-by-side box plot also effectively shows the variation in price for each car size via the interquartile range.

    The hollow histograms plot effectively shows the most common price range for each size of car (for example, the most common price for small cars is far easier to discern in the hollow histograms plot than the side-by-side box plot).

    3. I don’t really understand when to use row proportions and when to use column proportions..

  56. 1. Because pie charts are a very elementary way of representing data and are often not very useful.
    2.The components that I find most useful are the error bars that show the standard deviations.
    3.How do you choose a way of displaying data that is the most appropriate for a certain set of data?

  57. 1. There’s only one pie chart in the textbook because the book states that “it is more difficult to compare group sizes in a pie chart than in a bar plot”.
    2. I like the fact that in the side-by-side box plot, you can easily see how the medians of each compare to each other as well as the ranges of data in general. In the hollow histograms, I like that you can easily compare peaks/modes in the data.
    3. In general, the mosaic plot was a little harder to grasp at first because it was the only method used in the section that I had not seen previously. However, I do understand it more now after taking a little more time to comprehend it. I don’t have any specific questions.

  58. 1. Because pie charts make comparing group size more difficult than a bar plot. Pie charts are not as useful at displaying information graphically.

    2. The box-plot makes finding the interquartile range (median 50% of data points) and median very easy. It also displayed outlying data points nicely. While the hollow histogram displayed the relative “density” of the cars data at each price. This made identifying the relative number of cars in each price range simple.

    3. The Box-plot in figure 1.35. Is this also sometimes referred to as a box and whisker plot?

  59. Other types of charts are usually a more effective means of data visualization.

    I like the way the hollow histograms on the right panel overlap. This makes it easier to see where you budget falls in terms of the number and variety of options available for a given price.

    I don’t think I really understand the way the side-by-side box plot.

  60. 1.) It is more difficult to discern the proportions of different categories displayed on a pie chart than on a bar chart. For this reason, bar charts are the preferred method for displaying elements in various data categories via proportions.

    2.) I appreciate the power of the box-and-whisker plot to demonstrate outliers in the dataset. But I also see the value in the hollow histogram of the data. It displays a more detailed distribution of the data.

    3.) I am a little confused on how to visualize the data in a relative frequency table.

  61. 1. It seems to be that the reason for there only being 1 pie chart is that the pie chart is not a particularly useful means of displaying data. While appropriate at times, the same data could be better visualized using a bar plot instead

    2. For the side-by-side box plot, the most useful component is how the box is drawn. While initially looking at it, it can be difficult to understand (for the first time) the box plot displays average and max values well. The benefit of the histogram is that it is much more visual, and I’ve seen them more often. The simplicity of it is by far its most useful component.

    3. No real questions maybe just some clarification on the side-by-side box plot as the text does not really offer much

  62. 1) Pie charts are much harder to visualize and compare quantities on as opposed to bar charts.
    2) I thought the hollow histograms in the right panel were the most helpful because they allowed you to visualize the data distribution better.
    3)When would a pie chart be useful in visualizing data? Aesthetically I like them

  63. 1) Because although they’re well known, they don’t represent the data as well as other types of charts

    2) The advantage of the box plot is that you can see the exact range of prices for each type, and the advantage of the hollow histogram is that you can see how many of each price existed for each type.

    3) Can you explain box plots again.

  64. 1. There is only one pie chart in the textbook because pie charts do not always do a good job of displaying the relative proportions of the data they visualize. This is much better done by bar and mosaic plots.

    2. In the box plot the variability (error bar) was most useful, and in the hollow histograms the price axis was most useful. This is because in both visualizations they made it easy to see the category with the widest and smallest range of prices.

    3. Are pie charts really as deprecated as the text implies?

  65. 1. Pie charts are not as useful as other charts shown in the book. It is more difficult to compare group sizes with pie charts, as opposed to bar charts. For example, in Figure 1.33, the midsize and small sections look roughly the same size, but in the bar chart next to the pie chart, it’s obvious that the midsize region is larger.
    2. On the box plot, I find the spacing of the internal box to be useful as a visualization of the range that one variable may cover, and on the histogram, I find the effective y-axis to be the most useful as it enables me to visualize the difference in numbers between given variables.
    3. I’m a bit confused about the mosaic plots, especially Figure 1.31 (b) and 1.32. What are the advantages and disadvantages/differences between these two, and when would you be most likely to use them?

  66. 1. There’s only 1 pie chart in the textbook because it is more difficult to compare values that are very close in a pie chart.

    2. For the box plot, the most important component is the height of the boxes, it represents the price range for each type of car, whereas for the hollow histograms, the most important part is the height of the histograms, to know which price range has the most occurrence for different types of cars.

    3. For the box plot, what does the line in the boxes signify?

  67. 1) Pie charts are not as useful as other charts (such as bar graphs) in comparing the size of data because it only gives you percentages
    2) They say a lot about the distribution of the data (where the mean is, the greatest outliers, etc.) and also give a feel for how data from different categories match up with each other.
    3) Less of a question than a protest–I think pie charts are indispensable because they are so easy to understand. We used them many times in ENGM221 to represent market share of companies and they can be quite powerful in this application. I almost wonder if the book isn’t a little harsh in this respect?

  68. Why do you think that there’s only one pie chart in the textbook?
    What components of each plot in Figure 1.35 do you find most useful?
    What’s one question you have about the reading?

    1) Pie charts aren’t as useful as other kinds of visualization methods for examining and analyzing data. Relative proportions/sizes/quantities of “slices” within the pie chart are hard to distinguish without a numerical label on the slice, and then really the viewer is just comparing the number sizes, rather than the visualization method being useful.

    2) Side-by-side plot: while it’s pretty much the whole point of this type of plot, I find the fact that the box-and-whiskers plots are all next to each other on the same scale. It helps give these plots more meaning than they would alone, because you have the ability to compare them.
    Hollow histogram: I find the ability to see multiple histograms in a single data visualization to be the most useful. It lets one easily and quickly compare different histograms, either in their entirety or just at a particular area/section of the graph.

    3) The reading brings up an interesting point: pie graphs just aren’t that useful when one considers the other kinds of data visualization and comparison methods available. So why are they so extremely prevalent and taught so early in grade school?

  69. Why do you think that there’s only one pie chart in the textbook?
    – Pie charts are not particularly effective at communicating information. I think they only included one because pie charts are still in wide use despite their ineffectiveness. So, the authors would want their readers to be informed of what their use.
    What components of each plot in Figure 1.35 do you find most useful?
    – I find the variance bars most useful in the box plot chart.
    – I find the overlapping bins in the hollow histogram chart.
    What’s one question you have about the reading?
    – None.

Leave a Reply to Colin Cancel reply

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