# Run this cell first. Do NOT edit this cell.

Answer$1=$ Answer$2=$ Answer$3=$ Answer$4=$ Answer$5=$ Answer$6=$ None

import pandas

import numpy

import matplotlib

import matplotlib.pyplot as plt

#omatplotlib inline

from sklearn. linear$\_$model import LinearRegression

elect $=$ pandas. read$\_$csv$(\text{'}2020\_$US$\_$County$\_$Level$\_$Presidential$\_$Results.csv$\text{'})$

county $=$ pandas.read$\_$csv$(\text{'}$County$\_$demographics.csv$\text{'})$

state $=$ pandas. read$\_$csv$(\text{'}$StateNameData$.$csv$\text{'})$

elect.shape, county.shape, state.shape

$((3152,10),(3139,43),(51,3))$

Problem $1$

Interesting data science projects often combine data from multiple sources to investigate novel relationships. Process the above datasets to link the

per$\_$point$\_$diff in the $2020$ Presidential Election results to the Population.Population per Square Mile in the $2020$ census County

Demographics.

For each county, per$\_$point$\_$diff measures the differential in the percentage of votes recieved by the $2$ major parties $($as percent$\_$gop minus

percent$\_$dem$).$

To measure density, because of skew, we will use numpy. $log10($Population$.$Population per Square Mile$)$ and call it Log$\_$Pop$\_$SqMi.

In Answer$1,$ create a DataFrame with the following form, sorted by per$\_$point$\_$diff $.$ An example row is shown. You should be able to link the

variables for $3109$ counties using the given data; ignore any other counties.

Hints: StateNameData might help make the link. It is possible to join on multiple key columns.