Data Analysis with R
Starting up
Setting your library.
getwd()## [1] "C:/Users/tomas/Downloads/eda-course-materials"setwd("C://Users//tomas//Downloads//eda-course-materials")
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(ggplot2)
library(dslabs)
library(readr)
library(gridExtra)##
## Attaching package: 'gridExtra'## The following object is masked from 'package:dplyr':
##
## combinelibrary(reshape2)Accesing the data set.
pf <- read_tsv('C://Users//tomas//Downloads//eda-course-materials//lesson3//pseudo_facebook.tsv')##
## -- Column specification --------------------------------------------------------
## cols(
## userid = col_double(),
## age = col_double(),
## dob_day = col_double(),
## dob_year = col_double(),
## dob_month = col_double(),
## gender = col_character(),
## tenure = col_double(),
## friend_count = col_double(),
## friendships_initiated = col_double(),
## likes = col_double(),
## likes_received = col_double(),
## mobile_likes = col_double(),
## mobile_likes_received = col_double(),
## www_likes = col_double(),
## www_likes_received = col_double()
## )Scatterplots
To create a scatterplot, we add the x and y variables to qplot()
qplot(data=pf, x=age, y=friend_count)
ggplot() syntax is more specialized.
ggplot(data=DATAFRAME, aes(x=VARIABLE, y=VARIABLE)). aes stands for aesthetic, in aes you configure the color, size.
Then you can start adding layers:
geom_point() xlim(#,#)
ggplot(data=pf, aes(age, friend_count))+
geom_point() +
xlim(13,99)## Warning: Removed 4012 rows containing missing values (geom_point).
The alpha parameter inside the geom_FUNCTION() can set the transparency. Also geom_jitter() can give us a better view of many dots when they are grouped up.
ggplot(data=pf, aes(age, friend_count))+
geom_jitter(alpha=1/20) +
xlim(13,99) + scale_y_continuous(trans = "sqrt")## Warning: Removed 4284 rows containing missing values (geom_point).
dplyr
Using the pipe (%>%) we can modify dataframes and create new ones.
The group_by() creates a new variable that groups all of the data points, sumarize(NAME=FUNCTION) creates new variables applying functions, and %>% arrange() can organize the data frame by a variable.
geom_line connects the dots of geom_point.
friends_by_age_group <- pf %>% group_by(age) %>% summarise(friends_mean = mean(friend_count), friends_median = median(as.numeric(friend_count)), count=n()) %>% arrange(age)## `summarise()` ungrouping output (override with `.groups` argument) head(friends_by_age_group, 20)## # A tibble: 20 x 4
## age friends_mean friends_median count
## <dbl> <dbl> <dbl> <int>
## 1 13 165. 74 484
## 2 14 251. 132 1925
## 3 15 348. 161 2618
## 4 16 352. 172. 3086
## 5 17 350. 156 3283
## 6 18 331. 162 5196
## 7 19 334. 157 4391
## 8 20 283. 135 3769
## 9 21 236. 121 3671
## 10 22 211. 106 3032
## 11 23 203. 93 4404
## 12 24 186. 92 2827
## 13 25 131. 62 3641
## 14 26 144. 75 2815
## 15 27 134. 72 2240
## 16 28 126. 66 2364
## 17 29 121. 66 1936
## 18 30 115. 67.5 1716
## 19 31 118. 63 1694
## 20 32 114. 63 1443 ggplot(data = friends_by_age_group, aes(age, friends_mean)) + geom_line() +
scale_x_continuous(breaks = seq(0,99, 5), limits = c(13,99))## Warning: Removed 14 row(s) containing missing values (geom_path).
In order to overlay multiple plots together, you can add them as a layer.
stat=“summary” builds new variables into the plot, like the mean function.
geom_quantile(quantiles=PERCENTAGES, method=“rqss”, lambda=0.1) rqss for smoothing. This plot adds the quantiles of the y data.
library(quantreg)## Loading required package: SparseM##
## Attaching package: 'SparseM'## The following object is masked from 'package:base':
##
## backsolveggplot(data=pf, aes(age, friend_count))+
geom_jitter(alpha=1/20, color="orange") +
coord_trans(xlim=c(13,80), ylim=c(0,1000)) +
geom_line(stat = "summary", fun.y=mean_se(pf$friend_count)) +
geom_quantile(quantiles=c(0.1, 0.9), method="rqss", lambda=0.1,
lty="dashed", color="blue") +
geom_quantile(quantiles=.5, method="rqss", lambda=0.1, color="darkblue")## Warning: Ignoring unknown parameters: fun.y## No summary function supplied, defaulting to `mean_se()`## Smoothing formula not specified. Using: y ~ qss(x, lambda = 0.1)
## Smoothing formula not specified. Using: y ~ qss(x, lambda = 0.1)
To further divide the bin, we can create a new data frame with the age by months, and show the correlation with cor.test(VARIABLE 1, VARIABLE 2, method=“pearson, kendall, or spearman”).
The geom_smooth() adds a line with showing the trend of the data.
library(gridExtra)
pf$age_with_months=pf$age+((12-pf$dob_month)/12)
pf.fc_by_age_months<- pf %>% group_by(age_with_months) %>%
summarise(friend_count_mean=mean(friend_count), friend_count_median=median(friend_count), n=n()) %>%
arrange(age_with_months)## `summarise()` ungrouping output (override with `.groups` argument)p1<-ggplot(data=subset(pf.fc_by_age_months, age_with_months<71),
aes(age_with_months, friend_count_mean)) +
geom_line() +
geom_smooth()
p2<- ggplot(data=subset(pf, age<71),
aes(age, friend_count))+
geom_jitter(alpha=1/20, color="orange") +
coord_trans(ylim=c(0,500)) +
geom_line(stat = "summary", fun.y=mean_se(pf$friend_count)) +
geom_quantile(quantiles=c(0.1, 0.9), method="rqss", lambda=0.1,
lty="dashed", color="blue") +
geom_quantile(quantiles=.5, method="rqss", lambda=0.1, color="darkblue") +
geom_smooth()## Warning: Ignoring unknown parameters: fun.ygrid.arrange(p1,p2, ncol=1)## `geom_smooth()` using method = 'loess' and formula 'y ~ x'## No summary function supplied, defaulting to `mean_se()`## Smoothing formula not specified. Using: y ~ qss(x, lambda = 0.1)
## Smoothing formula not specified. Using: y ~ qss(x, lambda = 0.1)## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'
cor.test(pf$age, pf$friend_count)##
## Pearson's product-moment correlation
##
## data: pf$age and pf$friend_count
## t = -8.6268, df = 99001, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## -0.03363072 -0.02118189
## sample estimates:
## cor
## -0.02740737