In diegovalle/aire.zmvm: Download Mexico City Pollution, Wind, and Temperature Data

knitr::opts_chunk$set(fig.width = 7, 
                      fig.height = 7,
                      fig.align = 'center',
                      out.width='95%', 
                      dpi = 200,
                      collapse = TRUE,
                      comment = "#>")

Because of low temperatures education authorities decided to suspend classes in public and private schools in some parts of Mexico City on Wednesday January 31, 2018. We will download and graph the daily morning hourly temperature and compare the days of January 30 and 31 with the rest of the month.

Data Acquisition

We use the get_station_month_data function to download hourly temperature data for the month of January.

## Auto-install required R packages
if (!require("pacman")) install.packages("pacman")
pacman::p_load(aire.zmvm, dplyr, ggplot2, lubridate)

# Download hourky (HORARIOS) temperature (TMP) data for January 2018
jan <- get_station_month_data("HORARIOS", "TMP", 2018, 1)

Here's what the data we downloaded looks like:

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

Data Cleaning

After acquiring the data all that's left to do is to clean the data. Note that since it is winter daylight saving time doesn't apply.

temp_morning <- jan %>%
  mutate(day = day(date)) %>%
  # Kids were on vaction until the 6th and only go to school during weekdays
  filter(hour %in% c(5:11) & day > 6 & wday(date)) %>%
  mutate(hour = factor(hour, levels = c(5:11)))

# The school suspension was announced on the afternoon of January 30
temp_morning$color <- "other school days"
temp_morning$color[which(temp_morning$date == "2018-01-30")] <- "Jan 30-31"
temp_morning$color[which(temp_morning$date == "2018-01-31")] <- "Jan 30-31"

# remove the stations that didn't report data during January 30-31
temp_morning <- temp_morning %>%
  group_by(station_code) %>%
  filter(sum(value[date == "2018-01-30"], na.rm = TRUE) > 0 &
           sum(value[date == "2018-01-31"], na.rm = TRUE) > 0) %>%
  ungroup() %>%
  na.omit()

# reorder the data so the facets with the lowest temperatures
# come first
temp_morning <- mutate(temp_morning,
                       station_code = reorder(station_code, 
                                              value, 
                                              min, 
                                              na.rm = TRUE))

Plot

ggplot(temp_morning, aes(hour, value, group = day, color = color)) +
  geom_line(alpha = .8, size = .7) +
  facet_wrap(~ station_code) +
  scale_colour_manual("day", values = c("#542788", "#fdb863")) +
  scale_x_discrete(breaks = c(5, 8, 11), 
                   labels = c("5:00 AM", "8:00 AM", "11:00 AM")) +
  xlab("hour") +
  scale_y_continuous(breaks = c(0, 10, 20), minor_breaks = c(5, 15)) +
  ylab(expression(paste("temperature [",degree,"C]"))) +
  ggtitle("Daily morning temperatures in Mexico City during January 2018, by station",
          subtitle = "Data only includes school days during January\nSource: SEDEMA") +
  theme_bw() + 
  theme(axis.text.x = element_text(angle = 80, hjust = 1),
        legend.position="top")