This actually doesn't work, because sometimes 1440 means 0 and sometimes it means 2400, depending on whether the delay crosses over from one day to the next.

where are the differences between the two pictures？？

I think it should be "any_of()"

"the value" has been repeated twice after "of the" sth is missing

the an "e" is missing

don't believe this + character serves any purpose. It allows 1 or more spaces to between the matching words of group 1 and 2. But why would there ever be more than 1 space?

Please note that the table "Masters" in Lahman is recalled in "People".

so (makes it clearer)

Should have a comma after this word

Behavior differs - the last values in each row are concatenated, so I get:

a b 1 23 4 56

Same general pattern for the other answers.

This error is addressed and corrected in 16.4.2 - maybe worth noting for readers who want to check out solution.

how about using proportions or percentages? Even plots would be more revealing with less cluttered points.

cancelled_prop = (cancelled_num / flights_num),

OR even better:

cancelled_prop = (cancelled_num / flights_num) * 100,

ggplot (cancelled_per_day, aes (x = cancelled_prop, y = avg_dep_delay)) + geom_point (color = 'blue') + geom_smooth (se = FALSE)

ggplot (cancelled_per_day, aes (x = cancelled_prop, y = avg_arr_delay)) + geom_point (color = 'red') + geom_smooth (se = FALSE, color = 'red')

I think this should be Feb 1 2003 (2003-02-01), based on: date_custom <- c("Day 01 Mon 02 Year 03", "Day 03 Mon 01 Year 01")

behavior is different now: "> read_csv("a,b\n\"1") Rows: 0 Columns: 2 <br /> -- Column specification --------------------------------------------------------------- Delimiter: "," chr (2): a, b

i Use spec() to retrieve the full column specification for this data. i Specify the column types or set show_col_types = FALSE to quiet this message.

A tibble: 0 x 2

... with 2 variables: a <chr>, b <chr>"

As above, behavior has changed and final two values now combined to "34".

read_csv behavior has changed, warning is now: "Warning message: One or more parsing issues, see problems() for details"

This is becase read_csv() and read_tsv() are special cases of the more general read_delim().

According to ?tibble::print.tbl we should use max_extra_cols now: "max_extra_cols Number of extra columns to print abbreviated information for, if the width is too small for the entire tibble. If NULL, the max_extra_cols option is used. The previously defined n_extra argument is soft-deprecated."

This should be air_time_delay_pct not air_time_delay. As it is, the tibble below shows the same results as above so we can't see the data sorted to show highest percent first.

Another way to select these columns is by excluding all other columns (either by name as below or column number):

select(flights, -(year:day), -sched_dep_time, -sched_arr_time, -(carrier:time_hour))

This isn't a good coding approach, but as an example it does demonstrate how to exclude columns which was one of the techniques used in this section of the text.

Day %d Mon %m Year %y

The tvhours variable is the hours watched per day. It's unlikely, that respondents actually watch TV 24 hours a day. Could be a misunderstanding of the question, or they could be answering figuratively (e.g. "oh, I watch TV 24/7").

Should be x%%5 here not x%%3

This is not quite true, it is just appearance due to choice of scale. If you calculate correlation coefficient between displ and hwy for each classes, it is smaller than 0.5 only for 2-seaters and minivans. P value for linear models is also larger than 0.05 only for 2-seaters and minivans.

Has this been updated to after_stat() instead of ..

We can add a new variable: price per carat. And we will see that it drops just before a "nice" carat weight and then sharply rises. So we can somehow try to include it in the model. Table and depth seem to influence price too.

I think this is a typo. It should be vice-versa "from 2 to 4" based on the code snippet below, shouldn't it?

In all cases of discrepancy, the error is exactly 24 hours. It means that the flight was postponed till next day but dep_time erroneously gives the same day.

Another, completely different approach is to calculate distribution parameters for each day (I used quartiles) and plot them against time.

The median very slowly declines over the year. Although the decline is slow linear regression analysis shows statistically significant correlation.

An easy way to do it without typing the entire alphabet:

alphabet <- letters

names(alphabet) <- LETTERS

Strictly speaking, this code replaces forward slashes with double backslashes. There seems to be a bug in R that prevents replacing with single slash.

I can only add that in most (2/3) of these 48 hours either wind speed OR wind gust OR visibility were worse than the mean wind speed or mean visibility respectively.

I would also filter out cancelled flights with filter(!is.na(arr_time)) Flights missing a tail number is filtered out automatically by count.

Funny enough, the condition that makes it most likely to see a delay is normal atmospheric pressure!

weather_delay_dep <- flights %>% select(dep_delay, origin, time_hour) %>% filter(!is.na(dep_delay)) %>% inner_join(weather) %>% filter(!is.na(pressure))

pressure_delay <- weather_delay_dep %>% mutate(pressure = round(pressure)) %>% group_by(pressure) %>% summarise(delay = median(dep_delay))

pressure_delay %>% ggplot(aes(x = pressure, y = delay))+ geom_line() + geom_point()

You can see it even before calculating median delay for each pressure value (I use median rather than mean because delay distributions are skewed):

weather_delay_dep %>% ggplot(aes(x = pressure, y = dep_delay)) + geom_point() + geom_smooth()

What happened in 1995? Global number of cases increased by two orders of magnitude.

Also, some countries have missing age groups in some years.

There are other missing years, not related to existence of the country. If your group by country followed by summarise(years = unique(year)) you will see that Albania has data for 1995 and 1997 but no data for 1996, Algeria has data for 1997 and 1999 but not for 1998, etc.

The dataset description says that it describes round diamonds so two dimensions should be almost identical. A good starting point could be plotting frequency distributions of all three dimensions together:

dimensions <- diamonds %>% pivot_longer(cols = c(x, y, z), names_to = 'dimension', values_to = 'value' )

dim_short <- dimensions %>% filter(value <= 10)

ggplot(data = dim_short, mapping = aes(x = value, colour = dimension)) + geom_freqpoly(binwidth = 0.01)

Lines for x and y are practically identical, while z is shifted to the left. It does not prove that these dimensions are equal for each diamond but is a good indicator.

number

duplicate word

"Refused"

Why isn't it mentioned as well?

just that

the

I don't get this match when passing the respective code to the console.

I don't get this match when passing the respective code to the console.

I don't get this match when passing the respective code to the console. Any ideas why? What does the s* do?

no closing curve here.

at least

should be tibble

outdated explanation. the last two columns were combined.

In the context of defining which flights are cancelled, shouldn't the dep_delay variable be considered more important here? As you go on to say, just because a flight as arr_delay as NA doesn't necessarily mean the flight was cancelled. Whereas every flight that has NA for dep_delay also has NA for dep_arr; these are the flights that have been cancelled.

Minor mistake:

The variable cty, city miles per gallon, is a continuous variable.

Another solution, that evades plotting the points twice, would be the following:

ggplot(mpg, aes(x = displ, y = hwy, fill = drv)) + geom_point(color = "white", shape = 21, size = 3, stroke = 2)

It should rather be as follows to properly pass on the parameter na.rm: m <- mean(x, na.rm = na.rm)

Shouldn't year in this dataset be categorical?

I guess the character vector should be string instead of nchar.

How does this work? How come the evaluation of the conditional y == (-)Inf happens in the index of y?

Why are there two backslashes replacing a single forward-slash?

I wonder if this was written early in the editing of the text. The main text already describes this. If this is here as a reminder, maybe it could be worded differently so that it doesn't seem as if it was written early in the editing of the text and sort of forgotten here?

Could this notation be explained in the main text so that when we try to solve the problem we have the tool at our disposal?

from ?after_stat

after_stat() replaces the old approaches of using either stat() or surrounding the variable names with ...

so the (new) solution would be:

ggplot(data = diamonds) + geom_bar(mapping = aes(x = cut, y = after_stat(count / max(count)), fill = color))

such as

Either the following table lists or the following tables list.

the plots

class needs to be with different fonts

the values

colors

facets only on one variable.

with

in the column dimension.

in the row dimension.

facets

the combinations

Is there a way of doing this more systematically through z-scores?

flights2 <- flights %>% left_join(weather) %>% pivot_longer(temp:visib, names_to = "weather_condition", values_to = "weather_value") %>% group_by(weather_condition) %>% mutate(weather_value = (weather_value - mean(weather_value, na.rm = TRUE)/sd(weather_value, na.rm = TRUE)))

flights2 %>% ggplot(aes(weather_value, dep_delay, fill = weather_condition)) + geom_point()

I should warn that, although I tried this, not only does it take long to render, but it also simply didn't color my points. I'm not sure why.

Is there a reason why Batting in particular is in the middle? Or could any of the other two have been placed there instead?

omission of write?

Suggested edit: we could write a FizzBuzz function

In the textbook, this is "~names", rather than "~key".

It's not really clear to me why doing the means here makes sense... I'll come back to this and maybe later it'll click, but so far I've re-read this three times in three different days and it still doesn't click... Maybe there's another way of phrasing why this is being done the way it is?

Could we use & too? Could that be written here?

I think the correct name of the function is sd()

Could this answer include when I would use it, which is what the question asks. I wasn't sure what to answer because I couldn't think of scenarios where a mutate wouldn't do the job just as well?

What is the antecedent of the word "it"??

affects?

I thought it is better to use X axis for variables of larger number of categories.

I agree with the statement and used that as a basis for my answer.

sent_with_words_end_s <- str_subset(sentences, "\b[A-Za-z]{3,}s\b") # Focusing on only those sentences that meets the specified criteria

words_end_s <- str_extract(sent_with_words_end_s, "\b[A-Za-z]{3,}s\b") # Words ending in s (contains both plural words like "planks" and non plural words like "Sickness"

str_view(words_end_s, "\b[A-Za-z]+[^s]s$") #Extracts only words that end in s but not in ss.

Even though this looks similar to the time in which we did "group = 1", this seems different enough to warrant an explanation? At least I don't really understand how this works. On a hunch, I tried replacing "factor(1)" with "identity", and I got the same graph. How did that work? I tried looking up factor in the documentation, and it clarified what that meant, but I still am not sure how that relates to ggplot.

Typo: it omitted.

Edit suggestion: Since (it) is always good practice to have clear

Can you explain this code to me? I've looked up the if_else function but I do not understand this code.

Why did you use filter arr_delay <= 0 and not arr_delay > 0 when we are looking for the plane with the worst on-time record? This sounds counterintuitive to me, what am I misunderstanding? Thank you.

For better clarity, change "this delay will not have those affects plans nor does it affect the total time spent traveling." to this delay will not affect those plans nor would it affect the total time spent traveling.

Thanks ever so much for this amazing set of solutions. I learned as much from this as from the R4DS textbook. I just made a donation to the kākāpō recovery (as suggested by the r4ds online version) and would be happy to make a similar donation to any charity of your choice. Just let me know what cause you'd like me to support. Thanks again for your awesome work!

1. Please can you explain to me why you included the argument, width = 1 for geom_bar? Without it, the pie doesn't look different. I believe you must have specified for a reason?

2. This was my attempt to answer the question. I'm not totally if the resulting plot makes much sense. Please take a look and let me know what you think. Thank you.

ggplot(diamonds, mapping = aes(x = cut, fill = color)) + geom_bar() + coord_polar()

Typo. I think should read, such as, not such.

Please, can you explain why you have dots in the code? Thanks.

Nothing important, but a couple of typos here: "Presdients" and "Roosevelth". The latter should be Eisenhower (as in the y-axis).

I guess an alternative would be to drop "arrange()" and use the slice_max() function instead? That is, slice_max(carat, n = 20, with_ties = FALSE).

We're asked to convert the lyrics to a function that can be generalised to any number of people in the bed. Isn't the following closer to answering the question fully:

library(english)

ten_bed <- function(x) { lyric_1 <- c("There were ") lyric_2 <- c(" in the bed And the little one said, “Roll over! Roll over!” So they all rolled over and one fell out\n") number <- as.character(as.english(x:1))

for (i in number) { if (i != "one") { cat(lyric_1) cat(i) cat(lyric_2) cat("\n") } else { cat("There was one in the bed And the little one said, “Alone at last!” “Good Night!”") } } }

ten_bed(x)

Confirm my hypothesis that flights that departed in minutes 20-30 and 50-60 are more likely to have departed early than on time or late. Hint: create a binary variable that tells you whether or not a flight departed early.

Here you counted the opposite, meaning the flights AFTER the first delay of greater than 60 minutes the code should've looked something like this:

flights %>% select(tailnum, year, month,day, dep_delay) %>% filter(!is.na(dep_delay)) %>% arrange(tailnum, year, month, day) %>% group_by(tailnum) %>% filter(dep_delay<=60 ) %>% summarise(total_flights = n()) %>% arrange(desc(total_flights))

NA

this should be arr_delay

To match the first expression, this should be group_by(dest)

Should be just "arriving"

There doesn't seem to be a solution for this

optionally drop + (both): str_split(x, ", (and )?")[[1]]

what pattern

Maybe the intention here was colour_match2 (with the knock on impact on the str_view_all(...) such that only the 2 relevant strings are shown, not 3)

Exercise 13.3.3 (final sub-task):

The Pitching table is mislabeled as Fielding in the diagram (though it is clear from the listed PKs)

Exercise 13.3.3 Questions have you compare 3 data sets, one of which is call "People" while in Answers the same is called "Master" (verified in R: identical(Master, People) == TRUE) Posted 2020-12-28

Several problems in this sentence. I think this is missing "if" before "vars"; "the value" is duplicated; "of the ," should probably be something like "of the variable vars".

Should be "the"

typo - duplicate "the value"

this also seems to work

str_subset(words, "(.)(.).*(\1\2)")

Again, Thank you for your time. Really helpfull solutions (They are double backslash, only one is shown though)

Concordance with var() breaks down if you introduce an NA to the vector. As written, the function will return NA if there an NA is in the vector no matter what na.rm is set to.

The commenter above asked if you need to add na.rm to the sum function. In that case it will return a number but not the same as var because length is still counting NA, which can't be overridden. E.g. If you have x <- c(1:10, NA), variance(x) will return 8.25, but var(x, na.rm = TRUE) will return 9.166667.

To fix concordance with NA values in vectors, I wrote the function as such:

variance <- function(x, na.rm = TRUE) {
    if (na.rm == TRUE) {
        x <- x[!is.na(x)]
    }
    n <- length(x)
    m <- mean(x)
    sq_err <- (x - m)^2
    sum(sq_err) / (n - 1)
}

Then if you use x <- c(1:10, NA) both variance(x, na.rm = TRUE) and var(x, na.rm = TRUE) return 9.166667 and both variance(x, na.rm = FALSE) and var(x, na.rm = FALSE) return NA.

*difference

air_time_delay_pct

two "flight"

I think this alternative answers Exercise 5.7.8 with fewer lines of code. Note though that the answers don't coincide, i.e., I don't havet any tailnum with zero flights before the first 1-hour delay:

flights %>%
  filter(!is.na(arr_delay)) %>%
  arrange(tailnum, arr_delay) %>%
  group_by(tailnum) %>%
  filter(cumall(arr_delay < 60)) %>%  ## all until this 
  condition is false
  summarise(n = n())

less than

I was under impression that we need to know how to code these 3 bullet points. Instead, I see only theoretically-philosophical answers.

It might be scale_colour_viridis_b() or scale_colour_viridirs_c().I use the form scale_colour_viridis() then it shows it is not availble.

When I cut the interval into more than 40 slices, I get that the graphic shows fluctuation in the interval of [0,2].When n = 50, it confused me that the delay down and then up in [1.2,1.4].And it may get the different graphic in the short intervals nearing zero when using different n .

when installing the package datamodelr Using R studio version 1.3.1093 and R version 4.0.3, I get the message: Warning in install.packages : package ‘datamodelr’ is not available for this version of R

Are these duplicated?

should be names_ptypes?

Doesn't one also need to add "na.rm = TRUE" here?

should be %m, since %m is about the month and %M is about minutes. Clearly it should be months here.

a typo, it should be "the" or "this"

Should not it be dep_delay_diff = dep_delay - dep_time_min - sched_dep_time_min ?

Exercise 12.2.3

table2 %>% pivot_wider(names_from = type, values_from=count) %>% group_by(year,country) %>% summarise(cases=sum(cases)) %>% ggplot(aes(x=year, y=cases,colour=country))+ geom_line()

I think this question should be rephrased. There is certainly a dmonths().

dmonths(1)

[1] "2629800s (~4.35 weeks)"

months(1)

[1] "1m 0d 0H 0M 0S"

Maybe dmonths() was added since the initial publication of this section? Whether it is makes sense to use is another question, and you cover that in the description. Perhaps it should ask, "What is the difference in output between months(1) and dmonths(1)? Is it logical to ever use dmonths()?" I'm sure someone will come up with a creative use!

in the code, you put panel.grid.minor = element_blank(). I'm having a hard time understanding how seemingly putting a NULL element in the minor grid can make it thinner but still exist?

In the text, one of the subtitles notes that "2seater" indicates sports cars, and that they are an exception to the negative trend due to their light weight (despite large engines). (I don't think this exercise was meant to refer to subcompact cars.) Are some cars double-listed as 2seater and subcompact ?

output

My RStudio R Markdown files also display output inside the editor. Not sure if this is due to an update

Shift

my suspicion that this exercise is outdated has now increased, since I get the warning The .drop argument of unnest() is deprecated as of tidyr 1.0.0. All list-columns are now preserved. whenever I try to use .drop, and the output is the same as in the provided answer whether I include it or not

Could you please explain this part too?

ie why did we transform year to have mean 0? My graph and model coefficient outputs look the same whether I include the - median(year) or not...

and is the resulting equation something like Intercept = coef[1]x^2 + coef[2]x?

this seems to be a typo since this sentence appears twice in the text

higher regression standard error

I get a number similar to this (but the opposite sign) if I do 2 ^ rsme, but a completely different number if I do the same summary calcuation with resid instead of lresid. Why is it mathematically valid to back transform a manipulated log residual, ie one that has been through sqrt() and similar functions, instead of back transforming the residuals (to either a ratio or actual distance) first?

Why is this method with lresid better than using resid?

Also below, what are the units for 23-31? Percent? Dollars?

Could you please show how you got this number? I tried to follow Exercise 24.2.2: I figured that 0.5 must be equal to r^a1, so I calculated that if a1 = 3.2, then the residual must correspond to 0.81 dollars. Obviously, that doesn't make sense, so I'm wondering what I did wrong.

Even trying to back transform the residual by exponentiating 2 ^ 0.5 yields 1.4, an also seemingly wrong number.

When I look at individual observations in the data and plug them into (y1/y0) = (x1/x0)^a1, I get 5.78 for a1. Also, individual observations line up with yours, which is that predictions seem to be ~$40 below the actual price.

It would be helpful if I knew the coefficients of the linear model, but when I use coef(mod), out comes like 10 coefficients, which seems unhelpful.

Update: I think I calculated resid wrong--I tried to back transform it using 2 ^ lresid, just like we did with lpred and lcarat, but I couldn't seem to get it... Is it because resid is calculated from a log minus a log? thus lresid actually represents the distance of the data from the model in a way that, bc of log rules, comes out to be resid = price / pred.....

Using this interpretation, which is my most confident one so far, I agree with the statement that a residual of +2 means the price was 4x lower than expected, but a residual of +/-0.5 would still not come out to be 40. +40% or -30%, maybe, but not +-40.

logb

x

not sure what this means, is there any way you could provide a visualization? or how could you avoid/address this problem?

delete

Why??

Needs library("hexbin") to work

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>%<br> summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

For example, "N103US" has 46 flights under 1(one) hour. However, 46 is the total number of flights of "N103US", and none of these flights exceeded 1 hour. So unless I'm wrong, we shouldn't include it (the aforementioned "N103US") as having 46 flights before the first flight that exceeds one hour, because flight number 47 (your next flight) may not exceed an hour either (considering extending the search to the year 2014 and so on)! It may never have delay of more than an hour... and if the scope is restricted to only 2013 the reasoning is the same, we should exclude the 677 flights without any delay more than one hour from the result presented.

test

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

... filter(mean(cumulative_hr_delays) == 0) %>%

summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

... filter(mean(cumulative_hr_delays) == 0) %>%

summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

... filter(mean(cumulative_hr_delays) == 0) %>%

summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We noticed that there are observations with no record of delay exceeding 1 hour!

... filter(mean(cumulative_hr_delays) == 0) %>%

summarise(n = n())

Results 677 "tailnum" without any record of delay greater than 1 hour!

We also notice cases where there are no records of delays longer than 1 hour!

..... filter(mean(cumulative_hr_delays) == 0) %>% summarise(n = n())

results: 677 "tailnum" no record of delay exceeding 1 hour!

should be %m

where did you get this number? the mean and median time for the first function divided by the second is closer to 80 than 325...

is "lengths()" a function? could not find it in help

latter

seems to be missing

what does this argument do if it's not used in the function?

do you mean str_length(pad)?

This is repeated twice.

5

na.rm = na.rm ?

jrnlod, your answer to this question is particularly great, thank you!!

I added

if (x == 0) { print(x) }

as I don't think 0 should be a fizz/buzz Is there a less manual way to do this?

should these parentheses contain operations like &, |, !, and "xor" rather than the logical comparisons like < and == ?

values

this question is difficult to understand bc it seems to employ circular logic: obviously flights that leave early are caused by flights that leave early. maybe it means to say that the hypothesis tests whether flights departing in minutes 20-30 and 50-60 are mostly early flights and not delayed ones.

is this exercise unfinished? there is no explanation for the results here. I tried filtering out the ones with differences divisible by 60 (%% 60 == 0) to account for discrepancies due to timezone but that didn't seem to help...

should mention whether this was our mistake in analysis or a data entry mistake that we can fix during analysis. it isn't clear in the text

Can you help me translate the command? I don't quite understand what the result of mean( x > y) represents.

[A-Za-z]+

can you show what this is? it isn't listed under ?str_view

this result is irrelevant to the question

Exercise 14.5.2 question 2 We have not learned 'unlist' yet, we learn that only in Ch 21

I got a list and ordered it using the following code, but it takes only the first word in each sentence (I guess), and str_extract_all doesn't work at all. Any other ideas?

tibble(word = (str_extract(sentences, boundary("word")))) %>% mutate(word = str_to_lower(word)) %>% count(word, sort = TRUE) %>% head(5)

A tibble: 5 x 2

word n

<chr> <int>

1 the 262

2 a 72

3 he 24

4 we 13

5 it 12

Exercise 12.4.3.2. We learn str_split only in the next section. My rather inelegant solution was: has_apo <- sentences %>% str_subset("\'") has_ap_sep <- has_apo %>% tibble(sentence = has_apo) %>% extract(sentence, c("before", "apo", "after"), "([A-Za-z]*)(\')([A-Za-z]+)", remove = FALSE)

cf https://en.wiktionary.org/wiki/Category:English_contractions There are some contractions where the apostrophe appears as the first character, so use * not + in the first [A-Za-z] expression

us

running that expression shows that? not sure what this is referring to

Instead of using year, month, day, hour, you can join on only 'origin' and 'time_hour'

how come the output is grouped even though you specified ungroup()? does arrange() implicitly discard duplicate results if the var to be arranged on is a result of summarizing? does this have something to do with the "summarize regrouping output" message, and if so why does summarize() have an effect on subsequent functions? I'm not sure this was mentioned elsewhere in the book.

For this data, we only have data from the year 2013. So why do we always group by year in the exercises? Isn't that redundant/useless?

, s

The result shows that this is in fact not a primary key, as n > 1 is very high. Why?

One of the variables is named 'n', so first rename that variable, then check:

babynames %>% rename(no = n) %>% count(year, sex, name, prop) %>% filter(n > 1) %>% nrow()

[1] 0

since the text says that inner join is almost never used for analysis, could you also explain when it would be used? for example, is the reason you're using it in all the exercises because it is a convenient way to drop NA observations?

I truncate

used

more than one sex?

some of the exercises for this subsection a little premature...for some readers (like me) it is difficult to understand exactly what's going on without being first introduced to basic content knowledge in keys.

This part doesn't seem to do what it says it does, although I might just be interpreting everything wrong. But it seems that you are trying to check whether every entry for a particular "country, year" combination would be missing (eg no data collected for that region that year). However, the section is described as checking if multiple columns have missing values (as compared to only one column for that observation having a missing value). At least, perhaps consider clarifying this section.

delete