In slava-kohut/R-Capstone-SoftDev: tools for working with the NOAA Significant Earthquakes dataset

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(eqr)
library(dplyr)
library(ggplot2)
data("eq_data")

Intro

The eqr package is a set of analytical tools designed for working with the NOAA Significant Earthquakes dataset. The workflow basically consists of the following steps:

• load and clean the data
• analyze the data using timeline plots and maps

In what follows, we will illustrate each of these steps by providing specific examples that feature eqr functions.

Data cleaning

This step is performed using the eq_clean_data. The user needs to provide a path to the data set and the corresponding file name:

eq_data <- eq_clean_data('./', 'noaa-data-set.txt')

As a result, one obtains a clean, ready-to-use data set that is compatible with plotting and mapping functions. The eq_clean_data includes a call to eq_location_clean whose goal is to extract a country name from the location string. The function is vectorized.

eq_location_clean('CANADA: OTTAWA')
eq_location_clean(c('CANADA: OTTAWA', 'COLOMBIA: BOGOTA'))

Mapping

Earthquake maps can be produced using the eq_map function.

eq_data %>% filter(COUNTRY == 'COLOMBIA') %>% eq_map()

These maps are fully interactive and are annotated using earthquake dates by default. Pointers' labels can be changed by supplying a column name that is used for generating a new set of labels (the annot_col parameter). Let's use the earthquake magnitude for annotations:

eq_data %>% filter(COUNTRY == 'COLOMBIA') %>% eq_map(annot_col = 'EQ_PRIMARY')

A more informative label (location + magnitude + the number of deaths) can be produced using the eq_create_label function:

eq_data %>% filter(COUNTRY == 'COLOMBIA' & DATE > as.Date('1990-01-01')) %>%
  mutate(labels = eq_create_label(COUNTRY, EQ_PRIMARY, DEATHS)) %>%
  eq_map(annot_col = 'labels')

Timeline plots

Timeline plots are a convenient way to visualize the earthquake sequential order along with other important information for a particular country. These plots are generated using two ggplot2 geom functions: (i) geom_timeline and (ii) geom_timeline_label. As an example, let's generate a timeline plot for earthquakes happened in Italy. The point size and color will be used for coding the earthquake magnitude and the number of deaths, respectively.

ggplot(data = eq_data %>% filter(COUNTRY == "ITALY"),
       aes(x = DATE, size = EQ_PRIMARY, color = TOTAL_DEATHS,
           xmin = as.Date('1950-01-01'),
           xmax = as.Date('2015-01-01'))) +
 geom_timeline() +
 theme_eq

Note that the data for more than one country can be plotted. Also, it is important to note that the input data can be pre-filtered using the xmin and xmax parameters. The number of output points' labels is controlled by the 'nmax' parameter which refers to the eathquake magnitude sorted in a descending order.

ggplot(data = eq_data %>% filter(COUNTRY %in% c('COLOMBIA', 'MEXICO', 'USA')),
       aes(x = DATE, y = COUNTRY, size = EQ_PRIMARY,
           color = TOTAL_DEATHS, xmin = as.Date('1970-01-01'),
           xmax = as.Date('2015-01-01'))) +
  geom_timeline_label(aes(label = as.character(DATE)), nmax = 2) +
  geom_timeline() +
  theme_eq

slava-kohut/R-Capstone-SoftDev documentation built on May 20, 2019, 10 a.m.