Purpose

I want to see how much vaccination effects the COVID outbreaks around the US.

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from sklearn.metrics import r2_score

df = pd.read_csv('data/covid_vax_2021_06_04.csv')
df.head()

def plot_vax(df, degree):
    coeffs = np.polyfit(df['Fully  Vaccinated'], df['cases per 100k'], degree)
    p = np.poly1d(coeffs)
    r2 = r2_score(df['cases per 100k'], p(df['Fully  Vaccinated']))

    plt.scatter(df['Fully  Vaccinated'], df['cases per 100k'], color='red')
    plt.plot(range(25,60), p(range(25,60)))
    plt.xlabel('Percent Fully Vaccinated')
    plt.ylabel('Cases per 100k people')
    plt.title(f'US State Cases vs Vaccinations. \n r-squared: {r2.round(3)}')

plot_vax(df, 1)

plot_vax(df, 2)

This one is clearly overfitting

plot_vax(df, 8)

	state	Cases Daily Avg.	cases per 100k	14-day change	Hospitalized Daily Avg.	Per 100,000	14-day change.1	Deaths Daily Avg.	Per 100,000.1	Fully Vaccinated
0	Vermont ›	8	1	–81%	5	<1	–57%	0.0	—	57
1	South Dakota ›	20	2	–51%	40	5	–34%	2.3	0.26	43
2	Nebraska ›	45	2	–42%	81	4	–25%	1.0	0.05	42
3	California ›	932	2	–29%	1359	3	–18%	26.1	0.07	44
4	New Hampshire ›	32	2	–71%	36	3	–40%	0.7	0.05	50