feedback_david.md

Feedback on Alt3

Author: David

General comments for all chapters:

• Spacing between comments: If there is a comment and right below the student enters the answer, I'd leave two lines empty so that once the student has filled in the answer there is still one empty line. Makes the final answer nicely formatted.
• Solution should always contain all the code, including sample code (according to Datacamp this prevents errors)

Chapter 1: Intro to R

General

• Formatting, space between object name and "<- "
• Lowercase variable names

Ex 1

• Does not accept 55,000,000 and 50,000,000 as values even though that is the number of books sold. Either accept or make clear that the unit is millions of books.

Ex 2

• Error in SCT, couldn't locate source. Did type the right answer.

Ex 4

• SCT not flexible enough. Didn't accept the answer area_g <- 1/2*pi*15^2. And the solution contradicts the instructions (does not store result in total_area)

Ex 5

• Specify speed. What unit? Miles per hours?

Chapter 2: Data-types

General

• Does not run at all in my teaching editor. I believe this is because there are no {r} tags in the file. Does the author have the same issue or is this problem localized to my Datacamp?

Chapter 3: Vectors

General

• Comments in sample_code are all in lowercase. I think beginnings of sentences should be uppercase to make it consistent with other exercises

Ex 1

• Backticks for code: c() should be c()

Ex 2

• This sentence does not make sense to me: "It's in the same order presented above: 257, 355, 243, 411, 288". Maybe change to "The calories values are: 257, etc"

Ex 3

• This sentence does not make sense: "We have successfully assigned the temperatures (numeric values) to temp and the city names(character values) to city. But can we associate the temperature to it's related city?". I assume it's referring to a different iteration of the exercise.
• Instructions: "(You can go back to the previous questions and copy the objects stored.)" I think this might be frustrating for some users. Why not just put desserts <- c("blabla", "bla") in the sample_code?
• I think the "It would look like this" paragraph on the left-hand side is not necessary, since it's duplicated in the sample_code

Ex 4

• Make it clear that the desserts object was carried over from previous exercise. Maybe add it to the first line of sample_code: desserts <- c("bla", "blabla"). Just to make it clear why it's class character.

Ex 5

• cost[3:5] Error: object 'cost' not found

• Instructions are ambiguous. First, I would not use the word "list", since that's a separate class in R, use vector. Second, the instructions sound like the user is supposed to show the first three elements of desserts, not calories.

Ex 6

• "In the previous question, we accessed the number of survivors for consequtive animals (1st three)." - This ex. is about food, so I assume it's an error. Also I would write "first" instead of 1st.
• sample_code: Replace the word "list" with vector or character vector
• No space between object name and brackets: cost [c(1,5)]

• cost[c(1,5)] Error: object 'cost' not found

• "But what if we want to access the number of survivors for any 2 specific animals. " Referring to different iteration and missing quotation mark
• "Isn`t...": Don't use backticks in the "Exercise completed" message, it changes to formatting to R code style

Ex 7

• "Exercise completed" message: Don't use backticks, changes formatting ("Let`s")

Ex 8

• Instructions: 18 should be 108?
• Unable to enter solution since I get an error in the console: "Error in run_code(dc$get("DC_Solution")) ... etc"

Chapter 4: Sorting

Ex 1

• Left hand side: "analyse". I think we may want to use US spelling ("analyze")
• "We're": Might be too informal? I personally would change to "We are" and stick to no-abbreviations in all exercises

• library(nmle) Error: there is no package called 'nmle'

• Variables list (weight, sex etc) is very hard to read, use bullet points?
• Instructions, Missing word: "store it the object"
• sample_code: Why all the code about states and murder rates? I think this may be confusing for some users without explicitly saying that this is a template to follow. Maybe move to left hand side of page?
• Confusing wording: "# Access weight from the dataset and store it in weight". Maybe change to "Access the weight column from the RatPupWeight dataset and store it in a new object called weight"
• Ambiguous: "Sort the object and save it in the same object". Descending or ascending. Sort by what?
• Accept min(weight) as alternative solution

Ex 2

• Error: there is no package called 'nmle'
• Left hand side is empty except for instructions. Use it to explain how order is different from sort.
• According to stye guidelines we should not use short object names like this, instead order or weight_order.
• I think the instructions are too helpful. They basically give the answer (o[1])

Ex 3

• Error in title of exercise? "3. New Codes" Instead should be something about the new command users are learning
• "Write one line of code that gives the index of the lowest population entry". What population data? Refers to different iteration?
• sample_code: Murders example is confusing (user doesn't know where it comes from and what it has to do with the weight data. And it doesn't run: > which.min(murders$total) # Error: object 'murders' not found
• The object weight is undefined, so one cannot complete this exercise without first defining it. Either make clear in the instructions that user has to define it, or preload it in sample code. The solution uses RatPupWeight$weight, which is confusing since it's different from the previous exercises where we defined a separate object (weight)
• Comment is ambiguous: "# Find the smallest value for weight". But really, what we are asking for is the index of the vector, not the value

Ex 4

• sample_code: "Define variable treat". How? The user doesn't know anything about the dataset, needs more detailed instructions
• Not obvious to me why we need a new object for treat, but we don't define a new object for weight (as we did in previous exercises)
• Instructions are too helpful, they directly give the answer: treat[which.min(RatPupWeight$weight)]

• library(nmle) Error: there is no package called 'nmle'

Ex 5

• R example is duplicated in left-hand side and sample_code. Looks a little cluttered. On left hand side each command should be in a new line
• I would copy treat <- RatPupWeight$Treatment over from the previous exercises. The instructions don't ask the user to create the treat object, so might be confusing.
• Error: there is no package called 'nmle'

Ex 6

• Does not run at all, object o undefined, error in console.
• Use self-explanatory name for object instead of o

Ex 7

• Add spacing: "7. NA"
• Backticks and no definite article: na_example represents...
• na_example is actually not loaded into the environment, so this exercise does not run properly

Ex 8

• Title might be too hard to understand for new R users
• What is the object ind? No explanation of what it contains and why it might be partially NA
• Accept alteranative solution: mean(na_example[!is.na(na_example)])

Chapter 5: Vector arithmetics

Ex 1

• Left hand side looks cluttered. Add line breaks between code.
• The sample_code is about temperatures, the instructions are about food. Mistake?
• Unclear what to remake? Had to use solution

Ex 2

• "We've" -> "We have"
• I got the right answer, but it was not accepted as a solution. Here is what I used: x <- 1:2500 sum(1/(x^3))
• Solution, formatting: No space between sum and parentheses sum (1/ x^3)

Ex 3

• In sample code, preload murders dataset and explain it's columns on the left hand side. Too little context to understand the instructions
• Use dslabs package
• Instruction is ambiguous, it asks for the average murder rate in the US. That number would be sum(murders$total$)/sum(murders$population$. But what the question really is asking for is to calculate a weighted mean, with weight=1 for each state.

Chapter 6: Indexing

Ex 1

• "We’re using the same dataset as we did in the chapter on Sorting - women." - I don't think this dataset was used in this iteration?
• Conflicting terminology in instructions and sample_code... One refers to trees, one to women
• Instructions: Why does it say "lower than its mean of 2?" The mean of the data is not 2, it's approximately 2.098. Change to just "lower than 2"?

Ex 2

• Also mistake in terminology (trees/women)
• "Store the weight_rate < 2 in low" basically gives the answer away. I'd change to: "Create a vector low that is TRUE if an observation has a weight_rate below 2
• Left hand side should explain what which does, i.e. that it creates a vector of TRUE/FALSE, which can then in turn be used to select entries of another vector

Ex 4

• Error in console: "Runtime error. Object low not found."
• "Coder" language in comments might be hard to understand for beginners: "# Store height=60in, with weight rates lower than 2 in ind" - change to "Store observations 60 in tall and have a weight_rate below 2 in the object ind"
• Mistake in instructions: "low & food$courses=="Deserts", would give us the deserts from the courses variable in the food dataset." Not quite true, it only gives the indices of that vector. To get the values, we have to combine this with the repsective object the the [] brackets.

Ex 5

• Error in console: "Runtime error. Object weight_rate not found."
• Sample code: "How many women are > avg ? " - Confusing phrasing. Comments should be simple and self-explanatory, particularly to R beginners.

Ex 6

• "Coder" language in comments might be hard to understand for beginners: "# Store height=60in, with weight rates lower than 2 in ind" - change to "Store observations that are less than 60 in and weight rates below 2 in the object ind"
• Show in sample code that the dataset was loaded in and in what dataframe it was stored. Use dslabs package
• "(remember that they are character vectors)" -> There is only one character vector. Change to "Remember that the abbreviations are of class "character").
• "# Match the abbs to the murders$abb and store in ind"
• I'd use a more transparent name for variable ind, perhaps index
• Explain the columns of the murders dataframe. Otherwise it's unclear where the state names are stored and the exercise is hard to finish

Ex 7

• Terminology: %in% is not a function, it's an operator

Ex 8

• sample_code: Annoying for user since they have to go back to the previous exercise to look up the entries of the abbs vector. Just add that first line to the code: abbs <- c("ND", "NE", "NM", "NN", "NY")
• The last step is not obvious: "# Names of abbs in ind". Make explicit that you would like the user to display the name of the one abbreviation in abbs that is not part of the murders dataframe.

Chapter 8: Basic plots

Ex 1

• Error: there is no package called 'nmle'
• Provide more detail about the content of the dataset. For example, in which column in RatPupWeight is the litter size stored? Otherwise not clear to user how to store those values
• Scatterplot: Either be explicit about which variable should be on X/Y axes, or accept alternative answer where lsize is on the X axis: plot(lsize, weight)
• I'd also accept the alternative solutions plot(weight ~ lsize) and plot(lsize ~ weight)

Ex 2

• Runtime error: there is no package called 'nmle', exercise does not run or accept solutions

Ex 3

• "# Store values as in previous questions" - forces user to go back, just fill it in for them: weight <- RatPupWeight$weight lsize <- RatPupWeight$Lsize
• Header of sample_code should contain the name of the dataframe we are working with so that users don't have to go back when they write data = RatPupWeights in the boxplot command
• Again, a runtime error due to package nmle, cannot submit solution

Chapter 9: Programing Basics

Ex 1

• Too easy? User can just copy code into console and look at the output without thinking about the structure of the code. Can we turn the console off?

Ex 2

• Cannot submit response, runtime error in console. I think there is some error in the SCT

Ex 3

• I really like this exercise, it's quite challenging. Maybe explain to user that they can use the <, > operators on character vectors just like on numeric vectors. So to select letters before and including M they can use <= "M". I didn't even know this.
• I would not use append, not clear to new users. Instead "add new_names as a new column to the dataframe murders"
• Can't submit answer, some error in SCT: Runtime error in console

Ex 4

• Is it possible in Datacamp to have the function run over several lines? When I tried it it didn't work, but I think it looks very cluttered in one line. I guess it's also good practice to format functions so that they go over several lines since that seems to be the norm in R.

Ex 5

• Hint: What is "rate_per_n(murders)"?
• My answer was not accepted as correct, not sure why: nplus_states <- function(x, n=100) { x[x$total > n,]$state }
• Actually, it looks like the SCT does not even accept the hardcoded solution as the answer. Error message: "Did your code produce the same output as rate_per_n(murders)?"

Ex 6

• Do you want the students to actually run the code in the console or just look at it and figure the answer out theoretically? If you want them to run the code, I think they may encounter the issue I mentioned above, i.e. that Datacamp does not seem to handle function statements that go over several lines, thus you can't just copy in the code from the left hand side
• I don't understand why the answer is x<-7. The functions don't change the global value of x, so shouldn't it still be 3?