sebastianbarfort / sds

Social Data Science, course at University of Copenhagen
http://sebastianbarfort.github.io/sds/
12 stars 17 forks source link

Project Description - Group 21 #68

Closed hallurp closed 8 years ago

hallurp commented 8 years ago

title: "Project Description group 21"

output: pdf_document

1 Our idea

What question are you trying to examine and why is it interesting? Has the question been studied before and in that case what is your contribution?

When Pope Francis visited the United States some weeks ago, he had one main message: the "global abolition" of the death penalty. He touched upon a debate brewing across the country. In the United States, capital punishment-also called the death penalty-is a legal sentence in 31 states. Since 1976, 1419 inmates have been executed in the US and from where 530 of these have been in Texas.^1

This has awaken our interest and we tried to find some datasets. We are aware that there are many articles and papers on the executions, but (from what we have seen) not in a particularly special way, in the meaning that the usual facts are presented in the usual way of data visualization. Our aim is to focus on the people who were executed by finding answers on the following questions: Where are they from? Which occupation did they have? How old were they? Which last words did they use? Did they show regret? But also: who were their victims?

2 Data

I want you to describe what kind of data you plan on using to answer the question, where the data is located and a brief outline of an approach to gathering the data. Do you plan on scraping it from a webpage, can the data be called from an API, etc.?

The data used to make this analysis is scraped from the website of the 'Texas Department of Criminal Justice"^2, which publishes information on every executed death row inmate since 1982. The information is divided up into three parts: the general ones are on the main-page itself, where we just scraped it. The more detailed offender-information are on a sub-page, which we could scrape by creating a loop. We used the same way to get the data about the inmate's last statements. To get even more information we merged this table with data, we found on the website of the "Death Penalty Information Center". ^3

The final data is then very detailed and contains information on their names, age when executed, date of execution, race, county, last statements and other offender information that could be interesting.

3 The method

what statistical method do you plan on using? Is this a prediction problem or are you trying to do inference, and what kind of statistical model do you expect to use?

We are thinking about doing some predictions on the inmates (e.g.: correlation between the time receiving the penalty to the execution, and the number of killed people; is there a difference between black/white/hispanic?). But these are just some thoughts we would like to experiment with. Is it maybe possible to do predictions on the next inmate or the next victim? Do people usually commit crimes against people of their own race/age/gender or is the opposite the case?

One possibility could be to predict the race/gender/age of the executed offender based on the words used in his/her last statement.