Assignment: Week 3

Synopsis

This R Markdown Notebook is my report for the Data Wrangling with R class assignment for Week 3.
The following report summarizes good practices when beginning work with a new data set.

Data Source

For this report we are looking at the daily average temperatures for the city of Cincinnati, Ohio from The University of Dayton archives.

The archive contains data from 1995-present and is updated on a regular basis.

Initial Findings

• File is of fixed-width type with white spaces as separator
• 14 missing values - should be removed or imputed
• NAs are coded as “-99”
• 2016 data is incomplete as the year is still running

Packages required

library('gdata')
library('ggplot2')
library('lubridate')
library('scales')

---

About the data

A brief summary of the data:

• Month: Integer representing the month for the observation
• Date: Integer representing provides the date for the observation
• Year: Integer representing provides the year for the observation
• Temperature: Numeric data representing average daily temperature in degree Fahrenhiet

---

Data description

Importing the data

url <- "http://academic.udayton.edu/kissock/http/Weather/gsod95-current/OHCINCIN.txt"
colNames <- c("Month", "Date", "Year", "Temperature")
cincy_weather <- read.table(url, col.names = colNames, na = "-99")

# number of rows
nrow(cincy_weather)

## [1] 7963

# number of columns
ncol(cincy_weather)

## [1] 4

# column names
names(cincy_weather)

## [1] "Month"       "Date"        "Year"        "Temperature"

# structure of data
str(cincy_weather, give.attr = F)

## 'data.frame':    7963 obs. of  4 variables:
##  $ Month      : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date       : int  1 2 3 4 5 6 7 8 9 10 ...
##  $ Year       : int  1995 1995 1995 1995 1995 1995 1995 1995 1995 1995 ...
##  $ Temperature: num  41.1 22.2 22.8 14.9 9.5 23.8 31.1 26.9 31.3 31.5 ...

# sample rows
head(cincy_weather)

# summary stats
summary(cincy_weather)

##      Month             Date            Year       Temperature   
##  Min.   : 1.000   Min.   : 1.00   Min.   :1995   Min.   :-2.20  
##  1st Qu.: 4.000   1st Qu.: 8.00   1st Qu.:2000   1st Qu.:40.20  
##  Median : 6.000   Median :16.00   Median :2005   Median :57.10  
##  Mean   : 6.479   Mean   :15.72   Mean   :2005   Mean   :54.73  
##  3rd Qu.: 9.000   3rd Qu.:23.00   3rd Qu.:2011   3rd Qu.:70.70  
##  Max.   :12.000   Max.   :31.00   Max.   :2016   Max.   :89.20  
##                                                  NA's   :14

# missing values
sum(is.na(cincy_weather))

## [1] 14

We can see that all the missing values are under temperature, let us impute the missing temperatures with the average temperature for that week.

missing_temps <- which(is.na(cincy_weather$Temperature))
week_average  <- unlist(lapply(missing_temps, function(x){
  week_start <- max(c(0, x-3))
  week_end   <- min(c(x+3, nrow(cincy_weather)))
  imputed_temp <- mean(cincy_weather$Temperature[week_start:week_end], na.rm = T)
  return(imputed_temp)
}))

cincy_weather$Temperature[missing_temps] <- week_average 
summary(cincy_weather)

##      Month             Date            Year       Temperature   
##  Min.   : 1.000   Min.   : 1.00   Min.   :1995   Min.   :-2.20  
##  1st Qu.: 4.000   1st Qu.: 8.00   1st Qu.:2000   1st Qu.:40.20  
##  Median : 6.000   Median :16.00   Median :2005   Median :57.10  
##  Mean   : 6.479   Mean   :15.72   Mean   :2005   Mean   :54.73  
##  3rd Qu.: 9.000   3rd Qu.:23.00   3rd Qu.:2011   3rd Qu.:70.75  
##  Max.   :12.000   Max.   :31.00   Max.   :2016   Max.   :89.20

---

Visualizations

complete_cincy_weather <- cincy_weather[cincy_weather$Year < 2016,]
mean_temp <- round(mean(complete_cincy_weather$Temperature),2)
ggplot(data = complete_cincy_weather) + 
  geom_line(mapping = aes(x = complete_cincy_weather$Year, y = complete_cincy_weather$Temperature),
            color = "#428bca",
            stat = "summary", fun.y = "mean") +
  geom_hline(mapping = aes(yintercept = mean_temp),
            color = "#5cb85c") +
  geom_text(aes(1995, mean_temp, label = mean_temp, vjust = -1), size = 3) +
  labs(x = "Year", y = "Temperature") +
  ggtitle('Average Yearly Temperature')

The above plot shows the change in yearly average temperature over the last 20 years. ***

complete_cincy_weather$Decade = ifelse(
  complete_cincy_weather$Year < 2000,
  1990,
  ifelse(complete_cincy_weather$Year < 2010, 2000, 2010)
)
complete_cincy_weather$Decade <- as.factor(complete_cincy_weather$Decade)

ggplot(data = complete_cincy_weather) +
  geom_boxplot(mapping = aes(x = as.factor(complete_cincy_weather$Year),
                         y = complete_cincy_weather$Temperature,
                         fill = complete_cincy_weather$Decade),
               outlier.shape = 1, outlier.size = 1
               ) +
  scale_fill_manual(values = c("#9068be", "#6ed3cf", "#e62739")) + 
  scale_x_discrete(breaks = seq(1995,2015,5)) +
  labs(x = "Decade", y = "Temperature") +
  ggtitle('Temperature Variation Across Decades') +
  guides(fill = guide_legend(title = "Decade", title.position = "top"))

The above plot shows the temparature variation across years and is colored by decade.

---

ggplot(data = complete_cincy_weather) +
  geom_bar(mapping = aes(x = as.factor(Month), y = Temperature),
           fill = "#6ed3cf",
           stat = "summary",
           fun.y = "mean") +
  scale_x_discrete(breaks=1:12,
                     labels=month.abb) +
  geom_hline(mapping = aes(yintercept = mean_temp),
             color = "#5cb85c") +
  geom_text(aes(1, mean_temp+2, label = mean_temp, vjust = -1), size = 2) +
  labs(x = "Month", y = "Temperature") +
  ggtitle('Temperature Variation Across Months') +
  coord_flip()

The above plot show the temperature variation between months over the last 20 years.