Why do people rebel? Opportunity

Instructions

  • For the data, right-click, “save/download file as…”
  • Open the data
  • Complete all of the tasks associated with that data and keep track of your answers / images
  • Once you are done, go on Canvas and answer the questions related to the assignment
    • the Canvas assignment is timed; if you open it prior to completing the tasks you will run out of time

Skills used:

  • filtering data
  • summary statistics (note: towards the end I discuss “dummy” [0,1] variables)
  • grouped summary, statistic
  • plotting

My awkward Google Sheets video tutorials are here

Warning

You must complete all of the tasks here BEFORE opening the Canvas quiz; otherwise, you will run out of time!

Task 1: The geography of war

You’re going to work with a random sample from the Uppsala Conflict Data Program at The Department of Peace and Conflict Research, Uppsala University in Oslo, Norway. The dataset is called the Geo-referenced Event Dataset (GED) which provides detailed geo-referenced information on violent events around the world.

Variable Description
id Unique event identifer
year Year event began
month Month event began
day Day event began
country Country of event
type_of_violence UCDP violence type
conflict_name Name of conflict
side_a Side A
side_b Side B
source_article Article event came from
latitude Latitude
longitude Longitude
low_death_estimate Low estimate of number of deaths
high_death_estimate High estimate of number of deaths
best_death_estimate Best estimate

We’re going to use this geo-coded data to map all the violent events in a country in a year of your choosing. Here’s how to do this:

  • Look through the data a bit

  • filter the data to a year and country of your choosing, being careful that your selection does not result in too few rows (you want at least 100 rows)

  • select only these columns: type_of_violence, latitude, longitude, best_death_estimate

  • paste them into a new spreadsheet, then File –> Download that spreadsheet as a .csv file (in Excel you would “save as” -> file format -> .csv)

  • Go to this very ugly yet masterful website and on the left, upload your csv file under “File #1”

    • Look at the “data point options” section of that same page

    • set colorize this field to custom field and then write in “type_of_violence”

    • set the “resize using this field” option to “custom field” and then write in “best_death_estimate”

    • finally, bottom left, hit “draw the map”

    • Screencap the map so you can turn it in

  1. Spend some time looking at the map you have generated, you can also click on the points. It helps to have Google Maps open in a separate tab to get a better sense for what you are looking at. Now, reflect on what you are seeing. Which country did you pick? Where are these attacks concentrated? How do they relate to the country’s geography? Where are they in relation to the capital, to its borders? Are there patterns in the number of deaths, or the type of violence? Describe what you see in 2-3 paragraphs. (5 points)

Task 2: Rebel contraband

How do rebels fund their movements? You are going to work with the Rebel Contraband Dataset. Each observation in this dataset is a year in which a rebel group was engaged in conflict with the state. We refer to each row as a conflict-year.

Using the dataset, answer the following questions:

  1. Approximately what percent of entries in the dataset relied on extortion? (1 point)
  1. Calculate the percent of entries that relied on the extortion of alluvial diamonds, broken down by country. In which country was this funding strategy the most common? (1 point)
  1. How have coca smuggling trends changed over time? Use a pivot table to find the percent of entries with coca smuggling per year in the data. Plot the results as a line graph and screencap the image (you will upload). (1 point)

  1. How have coca smuggling trends changed over time? Briefly interpret the graph. (1 point)

  2. Do groups that abduct their members rely on smuggling at higher or lower rates than groups that don’t abduct? Calculate the percent of entries in the dataset that use smuggling, comparing groups that do and don’t rely on abduction. What is the difference in smuggling use between groups that use abduction and those that don’t? Report as a difference in percentage points (e.g., 35% - 20% = 15%). (1 point)

Variable Description
country Country
year Year
sidea Side A in conflict (UCDP)
sideb Side B in conflict (UCDP)
region Region of the world
extortion Did group use extortion? (1 = yes, 0 = no)
smuggling Did group use smuggling (1 = yes, 0 = no)
RESOURCE_ACTION Did group use ACTION to acquire RESOURCE? (1 = yes, 0 = no)
external_support Did group receive external support from foreign government? (1 = yes, 0 = no)
forcedrecruit Did group forcibly recruit members? (1 = yes, 0 = no)
abduction Did group abduct new members? (1 = yes, 0 = no)