milestone 4

[sukh2929]

1.create dependency diagram - should be in the project directory 2.create docker file - should be in the project directory 3.create docker image 4.update readme to add the link of dependency diagram 5.update readme to use your project with and without Docker

[rachelywong]

According to TA and peer feedback this is what we need to change for script5:

1. • explain why we used balanced log regression (explain when we talk about the model scores)
2. • remove scatterplot matrix?
3. • explain sampling (why we chose to use 1000)
4. • be more explicit about which features we ended up keeping and dropping
5. • fix conclusion about model scores and overall conclusion (especially part about logistic regression balanced performing average)
6. • talk more about figure 1 in results and discussion
7. • add why fit and score times, accuracy, and f1 scores were chosen as evaluation metrics
8. • grammar check

[jraza19]

edits to the read me after making the dockerfile include:

• add random as a dependency
• change the project proposal last line
• add the new docker code to the readme file
• create new dependency diagram

[rachelywong]

add sentence about certain sample we pick with random.seed and how data could be missing

[jraza19]

edit to eda2.py to address TA feedback:

Split the data into training (0.8) and testing (0.2)

diabetes_with_race = pd.read_csv(datafile[1]) train_df, test_df = train_test_split(diabetes_with_race, test_size=0.2, random_state=123) diabetes_subset = train_df