In the project scenario, the previous interns left behind a Google Colab notebook for a decision tree model to predict loan repayments.
For this homework, you will dig in to what they left behind to develop your understanding of what they did, generate your questions for things you don't understand and also start to think about how you might improve the model.
Go to Project 1 (Top menu bar: Projects --> Project 1)
Read the Project Write up
Open and Read the project writeup / scenario doc (google doc) - Project 1 - Lending Club
Look at the questions below and consider them as you start the next step.
Make note of how you need to work with your team and what you will turn in
Investigate the existing/left-behind work
Make a copy of the notebook that was "left behind" by the previous interns and systematically go through it (notebook is referenced in the doc and can be found in the project folder)
Get and Load the data
Read the text areas and try to follow its logic.
Run the code cells.
Use GEMINI to explain code sections, or update the code to test out your own theories or avenues of investigation.
Come to class to meet with your team with answers to the question below:
Colab Notebook/Data:
What is one-hot encoding? And why did the previous person use it?
What are your opinions about the features they chose to build the model on?
What is the accuracy/precision/recall of the previous intern's model? If you're going to improve the model what do you think a reasonable target for accuracy precision and recall are? Why do you think that? What's your insight?
Write down bring back to your team?
What is the TL;DR of this project as you understand it.
What's blocking you from starting? from understanding?
Copy the questions. Write your answers into a Google doc, and post the link to that doc using the HW 6 turnin form.