Midterm Project
September 28, 2020
This project uses Warren Wilson College data collected on the farm weather station. You can collect it yourself at https://rainwise.net/weather/wwcfarm, but you can only collect a months worth of data at a time. For your convenience, I downloaded all data from 2017-present, cleaned it up a bit, and placed all of the data in one huge file called weather.csv. It is posted on Moodle.
Go ahead and download into your working directory, read it into R with the read.csv command, and save it as the dataframe weather.temp. (It’s large, so it may take R Markdown a bit to process this command.)
setwd("/Users/ethan/Desktop/Jame's R Labs")
weather.temp<-read.csv("weather.csv")Go ahead and view the top 10 lines of the file to see what it looks like. What are the variables present?
head(weather.temp, 10)## X Time Temp.Avg Wind.Chill Hum.Avg Windspeed Gust Month Year
## 1 1 2017-04-01 00:00:00 52.7 52.7 72 6.2 6.2 4 2017
## 2 2 2017-04-01 00:01:00 52.7 52.7 72 6.2 6.2 4 2017
## 3 3 2017-04-01 00:02:00 52.7 52.7 73 4.7 6.2 4 2017
## 4 4 2017-04-01 00:03:00 52.7 52.7 72 3.7 3.7 4 2017
## 5 5 2017-04-01 00:04:00 52.7 52.7 72 3.7 3.7 4 2017
## 6 6 2017-04-01 00:05:00 52.7 52.7 72 3.7 3.7 4 2017
## 7 7 2017-04-01 00:06:00 52.7 52.7 72 3.7 3.7 4 2017
## 8 8 2017-04-01 00:07:00 52.6 52.6 72 3.2 3.7 4 2017
## 9 9 2017-04-01 00:08:00 52.5 52.5 72 6.0 10.5 4 2017
## 10 10 2017-04-01 00:09:00 52.5 52.5 72 9.5 9.5 4 2017- One of the annoyances of my data cleaning is that R put in an extra column with the name “X”. What is in that column? You decide that you don’t need the column X or “Hum.Avg”. Create a new dataframe called weather by deleting these two columns from weather.temp. Show the head of the dataframe to illustrate that it worked.
weather<- weather.temp[,c(2:length(weather.temp))]
head(weather.temp)## X Time Temp.Avg Wind.Chill Hum.Avg Windspeed Gust Month Year
## 1 1 2017-04-01 00:00:00 52.7 52.7 72 6.2 6.2 4 2017
## 2 2 2017-04-01 00:01:00 52.7 52.7 72 6.2 6.2 4 2017
## 3 3 2017-04-01 00:02:00 52.7 52.7 73 4.7 6.2 4 2017
## 4 4 2017-04-01 00:03:00 52.7 52.7 72 3.7 3.7 4 2017
## 5 5 2017-04-01 00:04:00 52.7 52.7 72 3.7 3.7 4 2017
## 6 6 2017-04-01 00:05:00 52.7 52.7 72 3.7 3.7 4 2017head(weather)## Time Temp.Avg Wind.Chill Hum.Avg Windspeed Gust Month Year
## 1 2017-04-01 00:00:00 52.7 52.7 72 6.2 6.2 4 2017
## 2 2017-04-01 00:01:00 52.7 52.7 72 6.2 6.2 4 2017
## 3 2017-04-01 00:02:00 52.7 52.7 73 4.7 6.2 4 2017
## 4 2017-04-01 00:03:00 52.7 52.7 72 3.7 3.7 4 2017
## 5 2017-04-01 00:04:00 52.7 52.7 72 3.7 3.7 4 2017
## 6 2017-04-01 00:05:00 52.7 52.7 72 3.7 3.7 4 2017One of the trickiest things to work with in datasets are timestamps. Conveniently, a package related to ggplot2 called lubridate makes working with dates easier. In the console, use the install.packages command to install the lubridate package. Then load the library as follows:
library("lubridate")##
## Attaching package: 'lubridate'## The following objects are masked from 'package:base':
##
## date, intersect, setdiff, unionSometimes it is useful to know the day, month, or year of a timestamp. Lubridate has those commands built in, so you can type, for instance,
month(weather[1,1])## [1] 4year(weather[1,1])## [1] 2017weather[1,1]## [1] "2017-04-01 00:00:00"Check that these give reasonable answers by looking at the [1,1] entry of weather.
- Use these ideas to find the day and minute the 1000th row this data set was recorded.
day(weather[1000,1])## [1] 1month(weather[1000,1])## [1] 4- Use the names command to change the names to: Time, Temp.in.Farenheit, Wind.Chill, Wind.Speed, Gust, Month, Year. Show the head of your file to illustrate your command worked.
names(weather)<- c("Time", "Temp.in.Farenheit", "Wind.Chill", "Humidity.Average", "Wind.Speed", "Gust", "Month", "Year")
head(weather)## Time Temp.in.Farenheit Wind.Chill Humidity.Average Wind.Speed
## 1 2017-04-01 00:00:00 52.7 52.7 72 6.2
## 2 2017-04-01 00:01:00 52.7 52.7 72 6.2
## 3 2017-04-01 00:02:00 52.7 52.7 73 4.7
## 4 2017-04-01 00:03:00 52.7 52.7 72 3.7
## 5 2017-04-01 00:04:00 52.7 52.7 72 3.7
## 6 2017-04-01 00:05:00 52.7 52.7 72 3.7
## Gust Month Year
## 1 6.2 4 2017
## 2 6.2 4 2017
## 3 6.2 4 2017
## 4 3.7 4 2017
## 5 3.7 4 2017
## 6 3.7 4 2017- Add another column to the weather dataframe “Temp.in.Celcius”, two different ways:
temp.in.Celcius<- c(1:length(weather[,1]))
weather$Temp.in.Celcius <- temp.in.Celcius
weather["Temp.in.Celcius"] <- temp.in.Celcius
head(weather)## Time Temp.in.Farenheit Wind.Chill Humidity.Average Wind.Speed
## 1 2017-04-01 00:00:00 52.7 52.7 72 6.2
## 2 2017-04-01 00:01:00 52.7 52.7 72 6.2
## 3 2017-04-01 00:02:00 52.7 52.7 73 4.7
## 4 2017-04-01 00:03:00 52.7 52.7 72 3.7
## 5 2017-04-01 00:04:00 52.7 52.7 72 3.7
## 6 2017-04-01 00:05:00 52.7 52.7 72 3.7
## Gust Month Year Temp.in.Celcius
## 1 6.2 4 2017 1
## 2 6.2 4 2017 2
## 3 6.2 4 2017 3
## 4 3.7 4 2017 4
## 5 3.7 4 2017 5
## 6 3.7 4 2017 6- Write a loop that runs through every row of weather and computes the degree in Celcius using the formula.
convertToCelsius<- function(x){
temp<- (x-32)*5/9
return(temp)
}for (i in 1:length(weather[,1])){
#weather[i, "Temp.in.Celcius"]<- convertToCelsius(weather[i, "Temp.in.Farenheit"])
convertToCelsius(weather[i, "Temp.in.Farenheit"])
}- Now add a column (Temp.in.Celcius.Alternative) by using the fact that R does vector operations really well (so in particular, column operations). Note: this is easier than writing a loop.
temp.in.Celcius<- convertToCelsius(weather["Temp.in.Farenheit"])
weather$Temp.in.Celcius <- temp.in.Celcius
weather["Temp.in.Celcius"] <- temp.in.Celcius
head(weather)## Time Temp.in.Farenheit Wind.Chill Humidity.Average Wind.Speed
## 1 2017-04-01 00:00:00 52.7 52.7 72 6.2
## 2 2017-04-01 00:01:00 52.7 52.7 72 6.2
## 3 2017-04-01 00:02:00 52.7 52.7 73 4.7
## 4 2017-04-01 00:03:00 52.7 52.7 72 3.7
## 5 2017-04-01 00:04:00 52.7 52.7 72 3.7
## 6 2017-04-01 00:05:00 52.7 52.7 72 3.7
## Gust Month Year Temp.in.Celcius
## 1 6.2 4 2017 11.5
## 2 6.2 4 2017 11.5
## 3 6.2 4 2017 11.5
## 4 3.7 4 2017 11.5
## 5 3.7 4 2017 11.5
## 6 3.7 4 2017 11.5Again, show the head of the file to illustrate these two steps worked.
- Suppose you want to compute the Summary (5 Number Summary, plus Mean) of the columns Temp.in.Farenheit, Wind.Speed, and Gust. Do those separately in the following:
summary(weather["Temp.in.Farenheit"])## Temp.in.Farenheit
## Min. : 2.80
## 1st Qu.:42.30
## Median :55.00
## Mean :54.18
## 3rd Qu.:66.20
## Max. :91.70summary(weather["Wind.Speed"])## Wind.Speed
## Min. : 0.000
## 1st Qu.: 0.000
## Median : 2.800
## Mean : 3.481
## 3rd Qu.: 5.600
## Max. :29.100summary(weather["Gust"])## Gust
## Min. : 0.000
## 1st Qu.: 0.000
## Median : 4.200
## Mean : 5.364
## 3rd Qu.: 8.500
## Max. :44.500- It might be nice to store these results in a 3x6 matrix, called weather.summary. Do so, and look up rownames() and colnames() so you can name the rows and columns for easy readability.
(note: you don’t need to fill the matrix entry by entry individually. Remember subsetting in a matrix–how do you select the first row?)
Make sure you display your resulting matrix with appropriate row and column names!
weather.summary<- matrix(data = c(1:18), nrow = 3, ncol= 6)
#strsplit(summary(weather["Temp.in.Farenheit"]), ":")
justTheNumbers<- function(aDataframeCol){
bigVector<- c("")
for (i in 1:length(summary(aDataframeCol))){
newVar<- strsplit(summary(aDataframeCol), ":")[[i]][2]
bigVector[i]<- as.numeric(newVar)
}
return(bigVector)
}
weather.summary[1,]<-justTheNumbers(weather["Temp.in.Farenheit"])
weather.summary[2,]<-justTheNumbers(weather["Wind.Speed"])
weather.summary[3,]<-justTheNumbers(weather["Gust"])
rownames(weather.summary)<- c("Temp.in.Farenheit", "Wind.Speed", "Gust")
colnames(weather.summary)<-c("Min", "1st Qu", "Median", "Mean", "3rd Qu","Max")
weather.summary## Min 1st Qu Median Mean 3rd Qu Max
## Temp.in.Farenheit "2.8" "42.3" "55" "54.18" "66.2" "91.7"
## Wind.Speed "0" "0" "2.8" "3.481" "5.6" "29.1"
## Gust "0" "0" "4.2" "5.364" "8.5" "44.5"- Suppose I want a summary of all temperatures (in Farenheit) that were recorded in December. Subset all December data, and give the summary of it.
decembers<-c(which(weather["Month"]==12))
summary(weather[decembers, "Temp.in.Farenheit"])## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 16.3 33.0 41.0 41.1 48.6 71.3- Finally, I’d like a visualization of the temperatures by month, and also by year. Use ggplot to create the best visualization you can using side by side violin plots, faceted by year. You’ll probably want to Google afew things: for instance, help rotating text on the x-axis.
###This code swaps month numbers for month names, then makes a new column.
NewMonthNames<-c(1:length(weather[,1]))
weather["MonthName"]<-NewMonthNames
someVar<- c(unique(weather[,"MonthName"]))
length(someVar)## [1] 1632801NewMonthNames[which(weather["Month"]==1)]<-"Jan"
weather["MonthName"]<-NewMonthNames
someVar<- c(unique(weather[,"MonthName"]))
length(someVar)## [1] 1464502NewMonthNames[which(weather["Month"]==2)]<-"Feb"
NewMonthNames[which(weather["Month"]==3)]<-"Mar"
NewMonthNames[which(weather["Month"]==4)]<-"Apr"
NewMonthNames[which(weather["Month"]==5)]<-"May"
NewMonthNames[which(weather["Month"]==6)]<-"Jun"
NewMonthNames[which(weather["Month"]==7)]<-"Jul"
NewMonthNames[which(weather["Month"]==8)]<-"Aug"
NewMonthNames[which(weather["Month"]==9)]<-"Sep"
NewMonthNames[which(weather["Month"]==10)]<-"Oct"
NewMonthNames[which(weather["Month"]==11)]<-"Nov"
NewMonthNames[which(weather["Month"]==12)]<-"Dec"
weather["MonthName"]<-NewMonthNames
someVar<- c(unique(weather[,"MonthName"]))
length(someVar)## [1] 12NewMonthNames<-c(1:length(weather[,1]))
weather["MonthName"]<-NewMonthNames
someVar<- c(unique(weather[,"MonthName"]))
length(someVar)## [1] 1632801NewMonthNames[which(weather["Month"]==1)]<-"January"
weather["MonthName"]<-NewMonthNames
someVar<- c(unique(weather[,"MonthName"]))
length(someVar)## [1] 1464502NewMonthNames[which(weather["Month"]==2)]<-"February"
NewMonthNames[which(weather["Month"]==3)]<-"March "
NewMonthNames[which(weather["Month"]==4)]<-"April "
NewMonthNames[which(weather["Month"]==5)]<-"May "
NewMonthNames[which(weather["Month"]==6)]<-"June "
NewMonthNames[which(weather["Month"]==7)]<-"July "
NewMonthNames[which(weather["Month"]==8)]<-"August "
NewMonthNames[which(weather["Month"]==9)]<-"September"
NewMonthNames[which(weather["Month"]==10)]<-"October "
NewMonthNames[which(weather["Month"]==11)]<-"November "
NewMonthNames[which(weather["Month"]==12)]<-"December "
weather["MonthName"]<-NewMonthNames
someVar<- c(unique(weather[,"MonthName"]))
length(someVar)## [1] 12library(ggplot2) ## Warning: package 'ggplot2' was built under R version 4.0.3ggplot(weather, aes(x = reorder(MonthName, Month), y = Temp.in.Farenheit)) + geom_violin() + facet_grid("Year") + labs(title ="Temperature Distribution by Month, Jan 2017 - Aug 2020", x = "Month", y = "Temperature in Farenheit")+ theme(axis.text.x = element_text(angle = -45, hjust = -0.2))