duttashi / learnr Goto Github PK

A collection of self curated notes to understand data visualization techniques.

This Q was originally asked on SO. I'm reproducing it here for referencing purpose:

Suppose a dataset has a factor column with values like;

> mydata                    
   question id           value
1         1  1      not likely
2         2  1      not likely
3         3  1      not likely
4         4  1      not likely
5         5  1 slightly likely
6         1  2     very likely
7         2  2 slightly likely
8         3  2 slightly likely
9         4  2      not likely
10        5  2     very likely

So how do I group the factor levels for variable value into say two levels ?

This question was originally asked on SO. Reproducing it here for reference purpose only.

A minimum example:

library(tidyverse)
col1 <- c("UK", "US", "UK", "US")
col2 <- c("Tech", "Social", "Social", "Tech")
col3 <- c("0-5years", "6-10years", "0-5years", "0-5years")
col4 <- 1:4
col5 <- 5:8

df <- data.frame(col1, col2, col3, col4, col5)

result1 <- df %>% 
  group_by(col1, col2) %>% 
  summarize(sum1 = sum(col4, col5))

result2 <- df %>% 
  group_by(col2, col3) %>% 
  summarize(sum1 = sum(col4, col5))

result3 <- df %>% 
  group_by(col1, col3) %>% 
  summarize(sum1 = sum(col4, col5))

Question: At times, an error message like this, `Error: package or namespace load failed for ‘somepackagename’ in loadNamespace(i, c(lib.loc, .libPaths()), versionCheck = vI[[i]]): there is no package called ‘somepackagename’ would pop up. How to solve this problem?

Last evening I decided to update the installed packages. During the update process, a prompt came up, do you want to install from sources the packages which need compilation r. I chose the option, yes and boy, this messed up everything in RStudio.

Lesson learnt: if such a message pop's up, choose No. See this post for reference.

Another problem was I had never set up the library path for the installed packages. So, I needed to set the library path. And finally, I needed to tweak the .Rprofile.site file. For windows OS it is located in, C:\Program Files\R\R-3.5.0\etc

> str(data_balanced)
'data.frame':	610 obs. of  10 variables:
 $ VisitRsrc  : int  11 22 91 90 41 64 25 61 25 80 ...
 $ raisedhands: int  2 20 90 80 27 62 8 7 15 20 ...

> ggplot(data = data_balanced, aes(x=VisitRsrc, y=raisedhands, fill=gender)) +
+   geom_boxplot()+
+   coord_flip()+
+   scale_fill_discrete(name="Gender")+
+   facet_grid(~Relation)
Warning messages:
1: position_dodge requires non-overlapping x intervals 
2: position_dodge requires non-overlapping x intervals 

Issue: "warning message: position_dodge requires non-overlapping x intervals", when plotting a box plot is generated. This warning is generated, when plotting continuous variables on both x and y- axis

Often in a data analysis project, there arises a need or requirement to replace values in either a single or multiple column. In this question, I will address this issue from two perspectives;

• using baseR
• using mutate from tidyverse package

Suppose a given dataframe contains missing values.

# inject random NA values in the mtcars dataset
mtcarsNew<- as.data.frame(lapply(mtcars, function(cc) 
  cc[ sample(c(TRUE, NA), prob = c(0.85, 0.15), size = length(cc), replace = TRUE)
      ]
  )
  )
# total missing values
R> sum(is.na(mtcarsNew))
[1] 44
R> colnames(is.na(mtcarsNew))
 [1] "mpg"  "cyl"  "disp" "hp"   "drat" "wt"   "qsec" "vs"   "am"   "gear" "carb"

How to summarize every column in a dataframe such that it ignores the missing values

Often in a data analysis project, there arises a need or requirement to rename the column name in either a single or multiple column. An example is given below;

# create a data frame
> df<- data.frame(sample(4,size = 4, replace = TRUE),
                sample(4,size = 4, replace = TRUE),
                sample(4,size = 4, replace = TRUE),
                sample(4,size = 4, replace = TRUE)
                )
# show the column names
> colnames(df)
[1] "sample.4..size...4..replace...TRUE."  
[2] "sample.4..size...4..replace...TRUE..1"
[3] "sample.4..size...4..replace...TRUE..2"
[4] "sample.4..size...4..replace...TRUE..3"

As you can see the column names suck.. Need to make them more meaningful. How to do this?

Yesterday, when I launched the RStudio on my computer, I was greeted with this error message; To the best of my knowledge I did not change anything. I'm running RStudio version 1.0.136 and R version 3.3.3

Rterm.exe - System Error. 
  The program can't start because libatk-1.0-0.dll is missing from your computer. 
  Try reinstalling the program to fix this problem. 

. Clicking the OK button on the error message, will not close it. And, RStudio will not work any longer.

When plotting missing data using the VIM package, I got the following warning message;
Warning message: In plot.aggr(res, ...) : not enough horizontal space to display frequencies

The dataframe looks like the following;

> str(dengue.train$week_start_date) Factor w/ 1049 levels "1990-04-30","1990-05-07",..: 1 2 3 4 5 6 7 8 9 10 ...

As we can see, the variable, week_start_date is read in Factor or String format. How do I change it numeric format?

This question was originally asked on SO

Question

How to collapse (or merge?) rows with the same identifier and retain the non-empty (here, any nonzero values) values in each column?

Data

df = data.frame(produce = c("apples","apples", "bananas","bananas"),
                grocery1=c(0,1,1,1),
                grocery2=c(1,0,1,1),
                grocery3=c(0,0,1,1))

Desired output

 shopping grocery1 grocery2 grocery3
1   apples        1        1        0
2  bananas        1        1        1

The inspiration for this question arose from this StackExchange post. In this post, it was asked how to treat categorical data for outlier detection?

I want to sort a data.frame by multiple columns. For example, with the data.frame below I would like to sort by column z (descending) then by column b (ascending):

dd <- data.frame(b = factor(c("Hi", "Med", "Hi", "Low"), 
      levels = c("Low", "Med", "Hi"), ordered = TRUE),
      x = c("A", "D", "A", "C"), y = c(8, 3, 9, 9),
      z = c(1, 1, 1, 2))
dd
    b x y z
1  Hi A 8 1
2 Med D 3 1
3  Hi A 9 1
4 Low C 9 2

This question was originally asked on SO

Question: Suppose there are n dataframes (in this case 3). How to extract variables which appear in all n dataframes?

Dataset

df1 <- structure(list(Variable = c("a", "g", "e"), Val = c(0.9, 0.3, 
0.1)), class = "data.frame", row.names = c(NA, -3L))

df2 <- structure(list(Variable = c("h", "a", "e"), Val = c(0.2, 0.7, 
0.9)), class = "data.frame", row.names = c(NA, -3L))

df3 <- structure(list(Variable = c("z", "a", "e"), Val = c(0.5, 0.7, 
0.9)), class = "data.frame", row.names = c(NA, -3L))

In data visualization often we need to ascertain the appropriate visualization type.