gmcirco
commented
2 years ago @CynthiaLteif Lots of good questions here:
- Let's look at the code for this question:
names %>% filter( Gender =="M" & Year >= 1946 & Year <= 1964 ) %>% arrange( desc( Count ) ) %>% distinct( Name, .keep_all=T ) %>% top_n( 10, Count ) %>% select( Name, Year, Count ) %>% pander()
There are lot of things happening here. arrange( desc( Count ) ) sorts the names from highest to lowest by their counts. distinct( Name, .keep_all=T ) gets the unique names PER YEAR. You will notice that the names are sorted by Year-Name, so we only want the unique name-year combination (for example, James in 1947). The .keep = T just tells R to retain all the columns because by default distinct will throw away any columns you don't specify.
-
You can use mutate, but in this case we want to combine the rows together. Mutate will give you a column with the counts, but it will be repeated over the groups you have in
group_by -
It is often good practice to
ungroupafter usinggroup_bybecause in R any subsequent functions on that dataframe will be grouped. We might not want to perform a group operation, soungroupingmakes sure that you are starting on a normal dataframe. -
That's correct. The tilde refers to specific variables that are being mapped to the visualization. You then add additional layers afterwards.
layer_lines()is probably the most appropriate one here, but you can see it would look different if you added a bar chart vialayer_bars(). -
The example in this section first finds subsets of names that meet certain criteria in
df1,df2, anddf3(corresponding to: " (1) They were popular when your grandmother was young. (2) They were unpopular when your parent were young. (3) They have recently become popular again." ) You are correct that they are the name-year combinations. The peak variable finds the value, for each name, that it reached its highest value that year.
CynthiaLteif
I have a lot of question in week 5’s second reading “Data recipes using pipes”. Apologies in advance for the long post.
In “building a recipe to identify the top 10 male names of baby boomers” To keep only the most popular year for each name, we used: distinct(Name, .keep_all = T). How is this this keep ONLY the MOST popular year? What does .keep_all = T OR F do?
In the alternative approach of the “building a recipe to identify the top 10 male names of baby boomers” To count the total number of men given each name during this period we used: dplyr: summarize(total = sum(count)). Could we have used the mutuate(total = sum(count)) function instead?
Why do we use ungroup() after group_by()?
In ggvis package a. Does stroke = ~ identify the variable we’re examining? b. Does layer_lines() identify the type of graph? If yes, are there other types that coule be used?
In the hipster.names section, I did not understand how the calculations were done in the summarize(total = sum(count), peak = max(count)) function. a. Is the total summing the count in df1, df2, and df3 for each name per year? b. Is the peak choosing from the count, the highest count belonging to a specific year?
Thank you in advance! @gmcirco