```{r setup, include=FALSE} knitr::opts_chunk$set(echo = TRUE) ``` ## Section 1: Vector manipulation For this section, you will be working on the following vector ```{r} L = c(3, 1, 1.3, 2, '2', c(2,8), TRUE, FALSE, FALSE, TRUE, 2, 3, 3, 9,2, "banana", "tree", 'bee', 'this is a long string') ``` ### Q1. Write the code to show the data type of the 6th element in the vector L (Note: R uses the 1-based index). ```{r} typeof(L[6]) ``` Notice how R forced every element to be character type. ### Q2. What is the data type of the 4th (2) and 5th element ('2') in the vector L respectively? ```{r} typeof(L[4]) ``` ```{r} typeof(L[5]) ``` ### Q3. Find the total number of elements in the vector L. ```{r} length(L) ``` Notice how R will flatten everything. The 6th item, originally was a vector, but when put into the vector L as an element, it is forced to be flat and the 2 elements in it were released out. ### Q4. Find the length (number of characters) of the last element in the vector L. ```{r} nchar(L[length(L)]) ``` In R, there is no direct reverse indexing like the L[-1] in Python. But we can access the last element throught L[length(L)]. ```{r} L[-1] ``` Also notice the meaning of L[-1] in R: it will remove (-) the first element in the vector and return the rest. ### Q5. Use vector slicing to access the 4 boolean values in the vector L. ```{r} L[c(8:11)] ``` ### Q6. Print the vector in reverse order without changing it. ```{r} rev(L) ``` ### Q7. Write code to calculate the sum of the first three numbers in the vector L. ```{r} sum(as.numeric(L[c(1:3)])) ``` First select the first three numbers by L[c(1:3)]. Then turn them into numbers by the as.numeric() function. Do the sum(). ### Q8. Write code to count the frequency of 3 in the vector L. ```{r} table(L) ``` table() function in R can get the frequency table of items in the vector. ### Q9. Add the following 2 elements to the vector L: “illini”, “spring”. ```{r} c(L,c("illini","spring")) ``` ### Q10. Change the last element in the vector L to “this is a short string”. ```{r} L[length(L)] = "this is a short string" ``` ## Section 2: Data Frame In the following session, we will deal with a data frame. You can access the CSV file in the following link: https://drmaggiezhang.com/files/ElonMuskTweets.csv This file contains all tweets posted by Elon Musk in the past 2 years from Jan/01/2021 to Jan/01/2023. I did the data collection using Twitter API. By the end of 1st module, you will also be able to do it as well. ### Q11. Read in the CSV file into a data frame format. How many rows and columns of this data frame? ```{r} df = read.csv("https://drmaggiezhang.com/files/ElonMuskTweets.csv",stringsAsFactors = F) ``` The stringAsFactors argument is to make sure when reading csv files, the strings won't be treated as a factor. ### Q12. Find the most liked tweet with the maximum like_count. What is the id of that tweet? ```{r} df[df$like_count == max(df$like_count),] ``` ### Q13. What is the average retweet number of Elon Musk’s tweets? ```{r} mean(df$retweet_count) ``` ### Q14. What is the total length of tweet text? ```{r} sum(nchar(df$text)) ``` ### Q15. Write a function that takes in 3 parameters as input: retweet_count, reply_count, and like_count. Calculate their average as the output. Name the function as CalculateEngagement ```{r} CalculateEngagement = function(retweet_count, reply_count, like_count){ avg = (retweet_count + reply_count + like_count)/3 return(avg) } ``` ### Q16. Using the CalculateEngagement functions you just wrote, create a new column in the data frame called engagement_index. ```{r} df$engagement_index = CalculateEngagement(df$retweet_count, df$reply_count, df$like_count) ``` ### Q17. Write another function that takes in 2 parameters: retweet_count, and reply_count. Calculate the reply-to-retweet ratio as the output (reply_count/retweet_count). Name the function as CalculateRatio. ```{r} CalculateRatio = function(retweet_count, reply_count){ ratio = reply_count/retweet_count return(ratio) } ``` ### Q18. Similarly, using the CalculateRatio functions you just wrote, create a new column in the data frame called ratio. Note: there are some tweets whose retweet_count is 0. Therefore you need to write a conditional statement. If the retweet_count is 0, set the ratio to 0. ```{r} ratio = c() for (i in c(1:nrow(df))) { if (df$retweet_count[i] != 0){ r = CalculateRatio(df$retweet_count[i], df$reply_count[i]) }else{ r = 0 } ratio = c(ratio,r) } ``` ```{r} df$ratio = ratio ``` ### Q19. Get the full vector of all hashtags that appeared in Elon Musk’s tweets. (Note: you might need to combine the for loop and conditional statement, as well as the vector combination. The logic goes as follows: create an empty vector (named all_hashtags) to store all hashtags. For each item in the ElonMuskTweets[‘hashtags’], if it is empty, it means there is a vector of hashtags in that tweet. Append this vector to the all_hashtags vector you created. Use pandas.isna() to see if an element is empty). ```{r} all_hashtags = c() for (hashtag in df$hashtags) { if (nchar(hashtag) != 0){ all_hashtags = c(all_hashtags,hashtag) } } ``` ```{r} all_hashtags ``` ### Q20. Save the current data frame (with 2 additional columns) as a new CSV file. Upload it to Canvas. ```{r} write.csv(df,"results.csv") ``` 