• What does your data represent in the real world?
• How is this real world phenomena characterized by the data that you have?
• From what time period is the data?
• What's the source?

• Once you have this basic understanding of your data you can dig deeper.

• You can use visualization techniques to explore your data and derive some basic understandings of the phenomena you are studying, such as the largest and smallest values for each variable.

• Calculating summary statistics can translate the data into information by revealing the shape of the data, the mean, median, minimum value, maximum value, and variability.

For any data science project there are few simple steps to follow.

• Begin by creating a new folder that contains your data.
• Then create a new project in RStudio.
• Set your working directory to the folder you created above.
  
• Create a new RMarkdown document.
• Save it in your working directory.

• Go to file > New file > RMarkdown
• Save the file in you working directory
• Name it internet.Rmd

Planning your programs, presenting your code, and sharing your work.

• Create .Rmd
• Write text
• Embed code
• Render output

Resource: https://bookdown.org/yihui/rmarkdown/

---
title: "Hello R Markdown"
author: "Kristen Sosulski"
date: "2020-08-17"
output: html_document
---

This is a paragraph in an R Markdown document.
Below is a sample code chunk:

#{r chunkname, echo=TRUE, eval=TRUE}

myformula <- (2+2)

• metadata - YAML header
• text - Text outside of code chunks
• code - R code chunks

YAML Ain't markup language

#title: "Hello R Markdown"
#author: "Kristen Sosulski"
#date: "2020-08-20"
#output: html_document

• # Header 1
• ## Header 2
• ### Header 3
• ** Bold **
• _Italics_

echo=TRUE

2+2

## [1] 4

echo=FALSE

## [1] 4

plot(attitude)

https://github.com/rstudio/cheatsheets/raw/master/rmarkdown-2.0.pdf

We have a couple different options.

• read_csv() function from the readr library
• read.csv() function from the utils library.

• Typically faster (~10x)
• Produce tibbles, they don’t convert character vectors to factors, use row names, or munge the column names.
• Base R functions inherit some behavior from your OS and environment variables, so import code that works on your computer might not work on someone else’s.
• Column names that are numbers convert from 2002 to 2002 vs. X2002.
• Column names that contain spaces (bad practice) convert from Country Name to "Country Name" vs. Country.Name

internet_baser <- read.csv("world_internet_usage.csv")
head(internet_baser, 5)

##     country X2000 X2001 X2002 X2003 X2004 X2005 X2006 X2007 X2008 X2009
## 1     China  1.78  2.64  4.60  6.20  7.30  8.52 10.52 16.00 22.60 28.90
## 2    Mexico  5.08  7.04 11.90 12.90 14.10 17.21 19.52 20.81 21.71 26.34
## 3    Panama  6.55  7.27  8.52  9.99 11.14 11.48 17.35 22.29 33.82 39.08
## 4   Senegal  0.40  0.98  1.01  2.10  4.39  4.79  5.61  7.70 10.60 14.50
## 5 Singapore 36.00 41.67 47.00 53.84 62.00 61.00 59.00 69.90 69.00 69.00
##   X2010 X2011 X2012
## 1 34.30 38.30 42.30
## 2 31.05 34.96 38.42
## 3 40.10 42.70 45.20
## 4 16.00 17.50 19.20
## 5 71.00 71.00 74.18

library(readr)
internet_readr <- read_csv("world_internet_usage.csv")
head(internet_readr, 5)

## # A tibble: 5 x 14
##   country `2000` `2001` `2002` `2003` `2004` `2005` `2006` `2007` `2008`
##   <chr>    <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>
## 1 China     1.78   2.64   4.6    6.2    7.3    8.52  10.5    16     22.6
## 2 Mexico    5.08   7.04  11.9   12.9   14.1   17.2   19.5    20.8   21.7
## 3 Panama    6.55   7.27   8.52   9.99  11.1   11.5   17.4    22.3   33.8
## 4 Senegal   0.4    0.98   1.01   2.1    4.39   4.79   5.61    7.7   10.6
## 5 Singap…  36     41.7   47     53.8   62     61     59      69.9   69  
## # … with 4 more variables: `2009` <dbl>, `2010` <dbl>, `2011` <dbl>,
## #   `2012` <dbl>

As best practice, we're going to use read_csv() function from the readr library to import the world_internet_usage.csv data as tibble data frame.

class(internet_readr)

## [1] "tbl_df"     "tbl"        "data.frame"

Take a look at your data.

knitr::kable(head(internet_readr, 10))

• What function can we use?

hist(internet_readr$`2012`)

hist(internet_readr$`2012`, breaks=8, 
     main="Internet usage for 2012", 
     col="magenta", xlab=" ", labels=TRUE)

hist(internet_readr$`2012`, breaks=4, 
     main="Internet usage for 2012", 
     col="magenta", xlab=" ", labels=TRUE)

par(mfrow=c(6,3))
hist(internet_readr$`2000`)
hist(internet_readr$`2001`)
hist(internet_readr$`2002`)
hist(internet_readr$`2003`)
hist(internet_readr$`2004`)
hist(internet_readr$`2005`)
hist(internet_readr$`2006`)
hist(internet_readr$`2007`)
hist(internet_readr$`2008`)
hist(internet_readr$`2009`)
hist(internet_readr$`2010`)
hist(internet_readr$`2011`)
hist(internet_readr$`2012`)

#Redo the histogram matrix 
#and add aesthetics

boxplot(internet_readr$`2012`, 
        main="Internet usage for 2012", 
        col="magenta", 
        xlab=paste("The median is:", median(internet_readr$`2012`)),
        frame.plot=FALSE, horizontal=TRUE, 
        border="dark blue")

#Build box plots for 2000 -2012 
#based on the example above using par(mfrow=c(INSERT NUMBER OF ROWS, NUMBER OF COLUMNS)) 
#Include aesthetics.

• Technically, we would use geom_col(), geom_bar() or geom_line().

• However, what variable would you map for the x-axis and the y-axis?

ggplot(internet_readr,aes(XVALUE, YVALUE)) 
  + geom_col()

ggplot(internet_readr,aes(X,Y)) 
  + geom_col()

## [1] "Wide format"

## [1] "Long format"

• The gather() function from the tidyr package to reshape a tibble from wide to long form.
• Note the use of the pipe %>% that passes the left hand side of the operator to the first argument of the right hand side of the operator.

tidy_internet_readr <- 
internet_readr %>%
gather(`2000`:`2012`, key="year", 
       value="usage")

• Use write_csv(file, path)

library(readr)
internet_readr <- read_csv("worldtidy.csv")
head(internet_readr, 5)

## # A tibble: 5 x 3
##   country    year usage
##   <chr>     <int> <dbl>
## 1 China      2000  1.78
## 2 Mexico     2000  5.08
## 3 Panama     2000  6.55
## 4 Senegal    2000  0.4 
## 5 Singapore  2000 36

• Let's create a time series line graph.
• Use the ggplot2() package

ggplot(data, aes(x,y,color, group)) + geom_line() + ... + ...

library(ggplot2)
#assignment ggplot call to a variable
line01 <- ggplot(tidy_internet_readr,
        aes(x=year,y=usage,color=country,
            group=country)) + geom_line()

line01

• Labels: +labs(title="", subtitle="",x="", y="", caption="")