# Import libraries and DataFrame
#
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from pandas import DataFrame, Series
# Read the data from pokeman.csv into a DataFrame using pandas read_csv()
# Print out the first 6 lines of data using .head
df = pd.read_csv('pokeman.csv')
df.head(6)
# print out the data types of all features using .dtypes (no parentheses)
df.dtypes
# print out the column names using .columns
df.columns
# Create a pandas Series for the feature Speed; print out type
speed = df.Speed
print ( type(speed))
speed
# Create a NumPy array for the feature Speed (use.values) ; print out type
speedarray = df.Speed.values
print( type(speedarray))
print(speedarray)
# Make 1D NumPy arrays from the features Attack and Defense and do a scatter plot
# using matplotlib
#
a = df.Attack.values
d = df.Defense.values
plt.plot(a,d, 'ro')
# Create a new DataFrame "df_mod" which is same as original but we drop "Type 2" feature; print out to check
df_mod = pd.read_csv('pokeman.csv')
df_mod = df_mod.drop(columns = ['Type 2'])
df_mod
# Import libraries and DataFrame
#
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from pandas import DataFrame, Series
import seaborn as sns
# Read the data into a DataFrame
# Print out the first 5 lines of data
df = pd.read_csv('pokeman.csv' )
df.head(5)
# Add a white grid to the background of Seaborn plots using set_style
sns.set_style ("whitegrid")
# Make a scatter plot using Seaborn's relplot of Defense statistics (y-axis)
# vs Attacks Stats
sns.relplot(x='Attack', y='Defense', data = df )
# Repeat plot in previous cell but use color to indicate Type 1 (hue = )
sns.relplot(x='Attack', y='Defense', data = df, hue = 'Type 1' )
# Make a category plot of Defense statistics vs Type 1 (non-numerical)
# Rotation labels on x-axis for readability using plt.xticks using plt.xticks(rotation=-45)
sns.catplot(x='Type 1',y='Defense',data = df )
plt.xticks(rotation=-45)
# Make a Bar graph of Defense statistics for Type 1
sns.barplot (x='Type 1', y='Defense', data= df)
# Make a violin plot of the Defense data for Type 1
sns.violinplot(x='Type 1', y='Defense', data= df)
# Repeat the plot in the previous cell but change palette to 'prism' and change size
plt.figure(figsize = (10,6) ) # make the plot a bit larger with matplotlib
sns.violinplot(x='Type 1', y='Defense', data= df, palette='prism')
# Overlaying plots - overlay violin plot of Defense with actual points
# To do this (1) increase figure size using ```plt.figure(figsize = (10,6) )```;
# (2) create violin plot and set inner = None to get rid of the bars inside violin plot;
# (3) rotate x-axis labels for readability;
# (4) create swarmplot for points and set ```color='k'``` to create the points in black;
# (5) add title "Defense Data for Type 1"
#
plt.figure(figsize = (10,6) )
sns.violinplot(x='Type 1', y='Defense', data= df, inner=None)
sns.swarmplot(x='Type 1', y='Defense', data = df, color='k' )
plt.xticks(rotation=-45)
plt.title ("Type 1 Defense")
# Import libraries
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from pandas import DataFrame, Series
import seaborn as sns
# Import LinearRegression function from scikit-learn
from sklearn.linear_model import LinearRegression
# Read in data from file insurance.csv and create a DataFrame; print out some lines
df = pd.read_csv('insurance.csv')
df.head(7)
# Set background grid for Seaborn plots
sns.set_style ("whitegrid")
# Create scatter plot of charges vs BMI with color indiciating whether patient is
# smoker or not
e = df.charges
f = df.bmi
sns.relplot(e,f, data = df, hue = 'smoker')
# Get data to use for linear regression
# Right now we see if there is a relationship between insurance charges and bmi
bmi = df.bmi.values
charges = df.charges.values
# Make bmi an n by 1 array and charges n by 1
n = len(bmi)
bmi = np.reshape ( bmi, (n,1) )
charges = np.reshape ( charges, (n,1) )
# Create model and fit data
lr = LinearRegression()
lr.fit(bmi, charges)
# write out equation of line
print ( "intercept", lr.intercept_)
print ("slope", lr.coef_)
print ("The equation which fits the data in a linear regression sense is:")
print (f"{round (lr.intercept_[0],2)} + {round(lr.coef_[0,0],2)} times bmi")
# Use regplot to plot data and line
sns.regplot(x='bmi', y='charges', data=df)
# predict insurance costs for a person with BMI 31.7; round answer to nearest cent
bmi_eval = [31.7]
bmi_eval = np.reshape(bmi_eval, (1,1)) # requires a 2D array
pcharges=lr.predict(bmi_eval)
print(f"When the bmi is 31.7 the predicted charges are {round(pcharges[0,0],2)} dollars")
# Note that this value agrees with plot above because when x=31.7 y is around 14,000