Pick a city: Chennai, Hyderabad, Kolkata.

For your city, calculate the following: the minimum, maximum, mean, median, standard deviation, interquartile range, and the \(5^{th}\) and \(95^{th}\) percentiles (or quantiles).

Calculate the number of missing values by using functions sum and is.na.

We would like to find out what is the pollution cycle over a day for the city you chose. In the dataset, there is a variable, Time, which indicates the hour of the day at which the measurement was taken.

For your city, calculate the mean of the PM2.5 for each hour. Hint: use the aggregate function.

Use something like plot(Delhi_hour$Delhi) to explore the time series of measurements for your city. Use options such as

• type="l" to change from the default (points) to lines
• main="A great title" to provide a meaningful title
• xlab and ylab to produce human-friendly axis labels

We already collected the descriptive statistics for all the cities, and found that the overall maximum and minimum were approx 55.6 and 23.3.

• We should use a scale on the vertical axis that goes from 23 to 56.

• Use these values to customize the y-axis of the time series plot for Delhi.

• To do this, add an optional argument (named ylim) to the plot and assign it an object with these two values (described in the previous sentence) using c().

Pick any two cities – your team’s city plus one other (not Delhi).

On the lines immediately after the par command that you are given below, copy/paste the code for the two time series plots (with the standardized y-axis limits that you did above). Then run the little block of code. You should obtain two plots on the same graphical display, one above the other.

par(mfrow=c(2,1))
# YOUR WORK HERE for Challenge 4:
# plot() for the first city here...
# plot() for the second city here... resulting in two plots in the same graphic!
par(mfrow=c(1,1))

Reminder: This code merges two graphs, stacked one on top of another. Just like a data frame with 2 rows and 1 column, par(mfrow=c(1,2)) allows you to put two graphs side by side. When you finish you should always set the par back to par(mfrow=c(1,1)), otherwise your next graph will not have the right size.

In the previous challenge, we pick two cities and plot them in a multiple-plots. However, it was difficult to read due to the scale. To improve this, we can instead put both cities in one single graph. To do this, you first plot one city graph as you did in Challenge 4 but then immediately after you can use the lines function to add another city. You will have to assign a different color to distinguish the two cities.

Next let’s use the boxplot to gauge the distribution for each hour.

What is the pattern for the city?

What are the most important data presented in this plot?

Explain how the following code can be used to answer the question, and adapted to answer related questions.

Level <- 100
100*c(sum(x$Chennai >= Level, na.rm = TRUE)/sum(!is.na(x$Chennai)),
  sum(x$Delhi >= Level, na.rm = TRUE)/sum(!is.na(x$Delhi)),
  sum(x$Hyderabad >= Level, na.rm = TRUE)/sum(!is.na(x$Hyderabad)),
  sum(x$Kolkata >= Level, na.rm = TRUE)/sum(!is.na(x$Kolkata))
)

## [1] 5.8655222 1.2500000 0.1547988 0.1412429