For Design Exercise 9: Visualization Hand-Ins one of your “stories” must be built from a “class question”.
Here are the “class questions” - in that they are for the entire class. Note that for each question, there is a basic question, but then suggestions for more richness/detail/nuance to be layered into an answer/story. You must pick one of these four for the set of visualizations you create for DE9.
Note that several of these require doing data joins of the samples with the population/household data set.
How well balanced are the demographics of the sample across multiple demographic variables?
Are the samples balanced (or similar to the population) across the key demographic variables (age, sex, region, education, income, etc.)? Are there sufficient samples in the different intersectional groups (e.g., college-educated females in the northeast) to look for patterns? Does this hold across years? Are other sampling variables (weekend vs. weekday) balanced?
The definition of “well-balanced” can vary (distribution similar to the population, sufficient in all groups, etc.). The question might be better phrased as “what is the distribution of demographics across groups (especially intersectional groups)” with a more specific task of allowing for the assessment of balance / sampling quality.
In order to compare whether the balance matches the overall population, you need the statistics of the overall population, which may be difficulty to get in the right form. If you find a good source, please post it on Piazza!
Good answers to this question will help the viewer understand how balanced the sample is, across multiple variables (and intersections of multiple variables). It may be difficult to consider intersections of more than 2 or 3 variables - this is the subject of a Design Exercise 10: Project Proposal topic.
How has the distribution of how people used their time changed during the pandemic (e.g. 2019 vs 2021)? Are certain groups more or less affected?
Are the expected impacts visible in the sample? Are there surprising changes in time usage? Are there groups that had different changes?
Good solutions should help the viewer see how time usage changed (e.g., to identify specific categories that went up or down a lot), but also to identify groups where the changes were different. (identify differences and find groups with differences in differences)
Warning: the amount of samples for 2019 and 2021 is not that large, so be careful about trying to divide things too finely.
How does time usage (the distribution) change as income levels rise? What do poorer/richer people do more or less of?
I might suspect that you need to divide up the data (or filter it) in order to see through some of the variance. For example, kids may differ differently, or patterns may be different weekend vs. week-days.
Note that the variable used for income changed over the course of time. There is both HEFAMINC and HUFAMINC. At glance, the balance of incomes seems a little unexpected (high incomes are over-represented in the sample). It is probably good to group the income level codes a bit.
Good answers would show differences (and similarities) in the distributions, and break these apart.
As people age, they tend to work less. What do they do instead? Does this vary across groups?
This is a more specific question. Seeing the age vs. work trend takes a little looking. And “working less” might mean not working at all (being unemployed/retired), and day of the week will matter.
Good answers will expose some richness in the story.
Categorized Ideas
Your “class question” must be one fo the four above. Here are some other notes that might help you in choosing other questions.
General statistics
student questions
- How does the survey age group look like? Verify if it follows a normal distribution.
- Is the ATUS dataset a good dataset? Is it unbiased, balanced (age, gender, race, education, and occupation), and Well-designed?
- What is the distribution of time people spend on different activities and how it has changed over the years?
- What is demographics of the participants, in terms of year, age, gender, race, education, and employment?
Children / Parenting / Family
student questions
- I liked the question about amount of time spent raising children and working. I think it is an interesting question on its own, but can also be extended into a lot of other interesting questions as part of the story.
- How does the parents’ activities get affected by the number of children? Is there any activity that gets highly compromised as the number of children increase.
- How do people’s time spent on different activities change with the number of children < 18 in the household?
- What activities do people give up as they start having children? Do they return to these activities as their children become self-sufficient?
- How does the distribution of time for various activities change as the number of children in household increases? Which activities see an increase in time share?
- Does house income level have an effect on the quality of time that parents spend with their children? For example, assume parents have a high-income job. They might choose to send their children to a caregiving center because they spend most of their time at work. Instead, they would put more energy into spending their time with their children when they are not at work. However, parents who have a low-income job, even though the time they spend on their main job would be the same, might not have enough time to spend with their children because they spend their spare time doing second jobs. Visualizations might give answers to this broad question.
Gender / Partners / Family
student questions
- I think a question related to “Do females spend more time in the household compared to men?” (the effect of children in the household, etc) would be a good choice because I saw similar questions came up a lot and there are a lot of variables across categories that you could look into
- I feel the compiled questions have a lot of submissions that center around comparing the time spent by men and women on household activities. We would like to see how the trend has changed over time and see if women predominate men in this category.
- Is there any income bias between males and females?
- Historically there are differences in the norms of which genders take care of children or family members, how different are the amount of time spent between genders on these different care tasks?
- Time-sharing between partners: if both the survey respondent and their spouse work full-time jobs, how do they adjust other parts of their life relative to respondents whose spouse is unemployed?"
- How do men’s relationships with their families evolve over time? They spend less time talking to family members on the phone than women. Where is this time being spent instead? Do men gradually drift away from their parents/older relatives and spend their time with their newly established families? Do women maintain these familial connections?
Pre-Post COVID (careful - small samples)
A number of questions ask “what changed in how people spent their time pre/post covid.” Sometimes the questions focus on specific groups.
student questions
- Since we seen a drastic shift in our lifestyles due to covid, one important visualization would be to compare the time commitment for different activities pre and post covid.
- “I believe we should do an analysis of time spent across different activities for pre and post covid data i.e for the years 2019 and 2021.
- We hope to see changes like spending more time on personal care or family time than before as covid helped all of us realize many things. Out of which important was definitely “nothing should be taken for granted. And Health is wealth”
- What are the effect of Covid on people’s daily life? Are there any ups and downs?
- The time spent on leisure time and work before pandemic and after pandemic.
- the affect of covid on the habits of college age kids
- Does the COVID pandemic change people’s time use? What increases and what decreases?
- How does Covid affect people’s daily life? I found that religious activities & traveling have significantly decrease during covid and telephone calls have a significant increase.
- I like the question of how people’s behavior/time usage changed as the pandemic started, and now how behavior is changing again as the pandemic is starting to subside. It feels very relevant right now, as Madison is starting to feel pretty normal again, and as the pandemic time period has started to show up in this data.
- What sort of travel times are workers willing to put up with? Has it changed post-pandemic?
Activities over time
student questions
- How has the way people socialize changed over time? I think this provides an interesting and challenging question because there are a number of factors to consider when asking it. There is sex, family size, wealth, age, and many more factors that may change this, and doing a simple chronological analysis may not be fully accurate. There may have been large cultural shifts in between the generation of those in their 40’s and those in their 20’s that impacts the way they socialize outside of just age differences. Additionally, going on a year by year basis, we are not looking at the same population, so other underlying factors can have impacts on the results."
- “Are humans becoming impatient over time? How many hours on average are people willing to spend waiting for services? How has this changed over the years? Do some services have steeper declines than others?”
Activities by Age
student questions
- From the ATUS exercise, I think everyone should look at how time spent playing different sports varies by age, since I think this is interesting, multivariate, and isn’t entirely obvious.
- How much time do people spend socializing and relaxing and how does that vary with the age group?
- Because I think exploring the relationship between schedule and age is worthwhile since it reflects the interest variation.
- One question that everyone should ask is what is the distribution of the time spent on each activity for the entire population across all years? I think this is important because this “average could serve as a baseline to compare any deviations with. For example, if I am interested in the how people spend their time differently before and after the start of pandemic, then the population average can serve as a baseline to measure the deviation.
- “The overall distribution of different top level activities, the number of people and how long they spend doing those activities. This should be done for age categories.
Age General
student questions
- What is the distribution between each different activities that people do across all ages? And, maybe distribution between gender as well.
- how are different age groups within a different demographic spending time, or maybe what is each groups most time spent on
- What’s the relationship between age and time usage?
Employed vs. Unemployed, Earnings
student questions
- I think the relationship between income and how people spend time at work, and education is something everyone should consider. To be more specific, do people with higher income spend more time at work or do they have any other quality? This can be assessed in many ways like money and work time, or money and education, and others.
- Do those with unemployed spouses tend to work more? Do they spend less time with their kids, or doing household activities? Do they make more on average than those with employed spouses?"
- How do high earners manage their time differently than low earners?
- Do people earn more if they work more? There is large inflation at the highest weekly income. The median income is 60k. Interestingly, the number of people with the highest income 287K is the same as the number of people at 60K. The distribution is bimodal which is weird. Another interesting thing appears. As the income increased, the number of women in that decreased which may suggest some gender discrimination in wages. But there might be other issues.
- Which activities will be sacrifice when work hours increase? This question involves exploring activities relevant to work or may be affected by work, also involves observing data across time series.
Commonalities
student questions
- What is the composition of a day for each respondent?
- After working on the ATUS design exercise and talking with my group on Wednesday in class I believe the question that should be asked is what percentage of everyone’s daily activities are the same. This would be calculated by looking at the main high level categories to determine if a person is spending a piece of their day on the same activities as all of the other survey participants. Then the activities that are spent by all individuals would have their time added up and divided by the total number of hours in a day (24). This may come out to a very small percentage or even resulting in 0.00%. This question can be expanded by looking at groups of people based on their personal information and once again determining what percentage of that group of people’s days are spent on the same main categories. This will hopefully at least generate a non-zero percentage as the more specific and smaller group of survey participants should hopefully have more in common that the entire group of survey participants."
Relationships between time usages
student questions
- How much time do people spend on health and self-care? Does it change with work-related time spent? From the exploratory analysis, it appears that when the average time from health and self-care increases, people spend much less time at work and socializing. Now it can happen that people spend less time on work and more time in self-care on weekends and that caused the discussed trend. However, weekends should not stop people from socializing, and therefore, increased self-care and decreased socializing can be somewhat related. Maybe when people prioritize personal care more they cut time from socialization.
- Which activities will be sacrifice when work hours increase?
This question involves exploring activities relevant to work or may be affected by work, also involves observing data across time series. Also, this question is meaningful to explore"
- Do people who spend more time in religious and spiritual activities have a tendency to help non-household members and children? Do spiritual activities increase with age? We have always heard that religious people tend to help people, and it’s interesting to know if religious people would help their own family members in the same or different way they help strangers.
- How do recreational activities impact religious tendencies in young people? Is someone who spends more time meditating more likely to be religious than someone from the same age group who spends most of their time on the internet? Is internet usage connected to decrease in religious activity?
Specific Observations / Questions
student questions
- If we break down the population by sex and whether having children, there emerges an interesting pattern. First, for men who do not have children, the working time seems to decline after the start of the pandemic, whereas for men with children, the working time seems to increase. Second, for women, regardless of whether they have children or not, the working time steadily increases from 2019 to 2021.
- I ran a bunch of comparisons comparing the breakdown of time spent on speaking on the telephone at different age groups (people over and below the age of 70 and 55.) In one case, I found that people under the age of 55 spent more time in total chatting on the phone than people over the age of 55, but people over the age of 70 spent more time on the phone than people under the age of 70. Clearly something happens in the range 55-70 that causes the shift for this specific activity. Also this seems to be specific to men - in case of women, older women spent more time on the phone on average, irrespective of the age at which we split the two groups. Do other activities show such variations with age and gender? Or is talking on the telephone, usually being a social or work related activity something of a special case?
- Travel times traditionally have been significantly higher on Fridays than other days. Why is this (it seems to be a combination of traveling for work and doing non-grocery-shopping-but-still-shopping travel)? Why has this changed in recent years?
Not sure about data
student questions
- Do life-long learners do better in life? Here “better” could be make more money or have more leisure time.
- Golf is considered an old person sport/post-retirement sport/rich-person sport. Can the ATUS data establish any correlation between age and hours spent playing golf? Does it contrast with trends for sports involving more exertion, such as basketball?
