Overview

Due by 11:59pm on November 8th.

We learned three ways of collecting data from the internet:

• Accessing data using packages that wrap APIs
• Running (basic) API queries “by hand”
• Web scraping

For the homework, you will

• Combine two existing datasets in a novel (and reproducible!) way
• Create a new dataset by using an API or web scraping

Accessing the hw08 repository

Go here and find your copy of the hw08 repository. It follows the naming convention hw08-<USERNAME>. Clone the repository to your computer.

Part 1: Exploring the gapminder data (even more)

We’ve examined the gapminder data quite a bit. One relationship we’ve looked at (or are about to) is the relationship between population and life expectancy.

For the assignment, I want you to replace population with population density and evaluate its relationship with average life expectancy. To do this:

1. Get the country information using geonames - remember there is a R package for this (see the lecture notes)
2. Merge gapminder and the country information from geonames
   • Use left_join from dplyr to merge the tables
   • Note that you cannot directly do this - gapminder writes the name of countries differently from geonames. To complete the merge, you need a unique key to match observations between the data frames
   • There is neat little package for R called countrycode that helps solve this problem. countrycode() takes as an input a country’s name in a specific format and outputs it using whatever format you specify.
     • gapminder stores them using the country.name format
     • geonames stores them under the countryCode column using the iso2c format
     • I leave it to you to make the joining operation work
3. Calculate the population density for each observation
4. Produce an updated graph using population density
   • If you want to be real fancy, estimate a statistical model or compare the relationship across continents

Part 2: More complicated data collection

For the second part of the assignment, I want you to create a new dataset and analyze it. You can do so using any of the following methods:

• Install and play
• Write an API query function
• Scrape a website

If you go either of the last two routes, you need to write your own code or function to query the server and obtain the results. If you decide to skip the API entirely, you will need to use rvest to scrape the content of a web page and extract the relevant information.

If you use the install and play option, I expect immaculate and thorough analysis since you are choosing a much easier method to obtain your data. Consider yourself warned.

The end result must be a tidy data frame stored in the repository with some analytical component. This can be exploratory description and visualization or some method of statistical learning. I should be able to run your code and reproduce your data and analysis.^{1 2}

Submit the assignment

Your assignment should be submitted as a set of two Quarto documents using the gfm (GitHub Flavored Markdown) format. Whatever is necessary to show your code and present your results. Follow instructions on homework workflow.

Rubric

Needs improvement: Cannot get code to run. Fail to accurately create the population density variable. Generated data set is not tidy. No documentation explaining how to use your API function or web scraping script.

Satisfactory: Solid effort. Hits all the elements. No clear mistakes. Easy to follow (both the code and the output). Nothing spectacular, either bad or good.

Excellent: Estimate a statistical model for the gapminder question, or do something beyond producing a graph. Implement an advanced statistical learning model or extensive exploratory analysis. Write an API function that uses authentication.

Acknowledgments

• This page is derived in part from “UBC STAT 545A and 547M”, licensed under the CC BY-NC 3.0 Creative Commons License.