Weather Events and Their Health and Economic Effects¶

filePath = "./data/stormdata.csv.bz2"
if(!file.exists(filePath)) {
    download.file("https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2", filePath)
}

#Unzipping the file contents
bunzip2(filePath, "./data/stormdata.csv", remove = FALSE, skip = TRUE)

#Viewing extracted files
str_view(list.files("./data"), "stormdata.+", match = TRUE) 

#Reading into R
Sys.time()
stormdata_table = read.csv("./data/stormdata.csv")
Sys.time()

head(stormdata_table)

#Adding factor variable for creating panel plots
tt = mutate(sum_df, factor_var = rep_len(1:20, length.out = nrow(sum_df)))
tt2 = mutate(mean_df, factor_var = rep_len(1:20, length.out = nrow(sum_df)))

#Plotting
histogram(~ fatalitiessum | as.factor(factor_var), data = tt, layout = c(4,1))
##This is incomprehensible to me
##A plan B is needed

#creating a dataframe of each variable and its maximum value
results = data.frame(
            events =  names(tt[, 2:5]) , 
            quantity = apply(tt[, 2:5], 2, max)
                    )
results2 = data.frame(
            events =  names(tt2[, 2:5]) , 
            quantity = apply(tt2[, 2:5], 2, max)
                    )

#Plotting
ggplot(data = results, aes(events, quantity)) +
    geom_histogram(stat = "identity")

ggplot(data = results2, aes(events, quantity)) +
    geom_histogram(stat = "identity")

#sorting by occurences
rbind(
        arrange(tt, desc(fatalitiessum)) %>% head(n = 1),
        arrange(tt, desc(injuriessum)) %>% head(n = 1),
        arrange(tt, desc(propertydamagesum)) %>% head(n = 1),
        arrange(tt, desc(cropdamagesum)) %>% head(n = 1)
    )
#This category TORNADOES, TSTM WIND, HAIL has the most fatalities

rbind(
        arrange(tt2, desc(fatalitiesmean)) %>% head(n = 1),
        arrange(tt2, desc(injuriesmean)) %>% head(n = 1),
        arrange(tt2, desc(propertydamagemean)) %>% head(n = 1),
        arrange(tt2, desc(cropdamagemean)) %>% head(n = 1)
    )

#reading events data in
events_df = read.table("./data/events.csv", sep = "|", header = TRUE)
events_df %>% head

#replacing each observation interword spaces with |
events_var = str_replace_all(events_df$Event.Name, "[ |/]", "|")

events_var

#Taking a random sample of the dataframe observations to reduce complexity
set.seed(100)
sb_st = sample(x = 1:nrow(min_stormdata),
                   size = 10000,
                   replace = TRUE
                  )
rand_stormdata = min_stormdata[sb_st, ]
rand_stormdata %>% head

has_rownames(subset_matrix)
subset_matrix2 = rownames_to_column(subset_matrix, var = "Effects")

#Moving identified observations into rows
tidy_ev_table =  subset_matrix2 %>% 
                pivot_longer(names(subset_matrix) , names_to = "event", values_to = "figures") %>%

                #Moving effect variables into columns
                pivot_wider(names_from = "Effects", values_from = "figures")

fin_tally_df = sapply(1:nrow(tidy_ev_table2), summ_func) %>% data.frame
fin_tally_df

#Tidying operations
fin_tally_df2 = rownames_to_column(fin_tally_df, var = "title") %>%
                        pivot_longer(X1:X29, names_to = "arbitrary", values_to = "realvalues") %>%
                        pivot_wider(names_from = title, values_from = realvalues, ) %>%
                        select(-(arbitrary)) %>%
                        apply(MARGIN = 2, unlist) %>%
                        as.data.frame %>%
                        mutate(totalfatalities = as.numeric(totalfatalities), totalinjuries = as.numeric(totalinjuries),
                                totalprop = as.numeric(totalprop), totalprop = as.numeric(totalprop)
                             )
fin_tally_df2 %>% head

#Plotting
##Property Damage
ggplot(data = fin_tally_df2, 
       aes(evs, totalprop)) +
    geom_histogram(stat = "identity") + 
    ##Rotating label ticks
    theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))

##Crop Damage
ggplot(data = fin_tally_df2, 
       aes(evs, totalcrop)) +
    geom_histogram(stat = "identity") + 
    ##Rotating label ticks
    theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))

##Injuries
ggplot(data = fin_tally_df2, 
       aes(evs, totalinjuries)) +
    geom_histogram(stat = "identity") + 
    ##Rotating label ticks
    theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))

##Fatalities
ggplot(data = fin_tally_df2, 
       aes(evs, totalfatalities)) +
    geom_histogram(stat = "identity") + 
    ##Rotating label ticks
    theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))