Authors: Paul Fentress, Garrett Layton
OBJECTIVE: The Data Science and Machine Learning Life Cycles are missing a key component which is an ethical reflection step. We wanted to create a Python package that allows users to easily print out and follow through our Data Ethics agreement. By making a widely accessible Data Ethics agreement that can be directly integrated into Python projects, we provide a tool that makes it easier for anyone to start reflecting on the ethical implications of their work with data. The hope is that by providing an easy-to-use tool, Data Ethics can become more popular and accessible, to begin re-writing the narrative that data science and data ethics are two separate studies. The tool should be useful for those who are well versed in Data Ethics, and those totally new to Data Ethics. The purpose of this tool is not to create a stamp of approval for ethically used data, rather it is to make data ethics more accessible.
Problem: UC Berkeley Data100 Project Life Cycle Without Ethics.
Solution: Edited Version Of Data100 Life Cycle With Ethics.
Note: Although the diagram above shows an ethical reflection being added between asking a question and obtaining the data, we realize that ethics should be considered at each part of the data science life cycle. Furthermore, data science projects do not always start from the same point, so the ethical reflection should be moved around accordingly to the specific project.
One of the easiest way to describe our tool, is to show how the tool works with a demo. Bellow is the README file on the project’s GitHub, which shows how users can install, import, and use the data ethics agreement.
<aside> ⚙️ Link to GitHub:
https://github.com/fentresspaul61B/data_justice
</aside>
The package is fairly easy to use, and by using PyPi, anyone in the world using Python should be able to import the package and use the agreement. Below shows a demo of the installation, and then what follows are the actual steps if you would like to try it yourself, on your own machine.
First, make sure you have the most recent version of pip.
# If you are working in Jupyter:
!python3 -m pip install --upgrade pip
# If you are working in Terminal:
python3 -m pip install --upgrade pip
Install the data_justice package with the commands below.
# If you are working in Jupyter:
!pip install -i <https://test.pypi.org/simple/> pg-ethics==0.0.1
# If you are working in the terminal:
pip install -i <https://test.pypi.org/simple/> pg-ethics==0.0.1
Once the package has been installed successfully, run the command below in your notebook or terminal.
from data_justice import data_justice
To print the generate the data_justice agreement run:
data_justice.generate_agreement()
Copy the agreement and paste it into a markdown cell.
Follow the steps in the agreement!