#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Tue Mar 27 14:08:22 2018

@author: teacher
"""

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import statsmodels.formula.api as smf   # import model library

# Clean data as in Lab 14.  We also need to skip the final two lines of the CSV file because they are empty.
fields = ['Major','Unemployment Rate','Underemployment Rate','Median Wage Early Career','Median Wage Mid-Career','Share with Graduate Degree']
labor = pd.read_csv("labor-market-for-recent-college-grads.csv",skiprows =13, skipfooter = 2,usecols=fields)

# Tell pandas the median wage columns are really numbers
# by getting rid of the , and telling Python the string is a float (decimal number)
labor['Median Wage Early Career'] = labor['Median Wage Early Career'].str.replace(",","").astype(float)
labor['Median Wage Mid-Career'] = labor['Median Wage Mid-Career'].str.replace(",","").astype(float)

# statsmodel library needs column names not to have spaces
# so we will rename the columns
labor.columns = ['Major','Unemployment','Underemployment','Early','Mid','Share_grad']

# Set up the model and fit it (estimate the parameters) using our data.
# For linear models, the formula is y ~ x
# ols = ordinary least squares regression
lm = smf.ols(formula = 'Mid ~ Early', data = labor).fit()
# Print out the parameter values 
print(lm.params)

# Make some predictions
# First we make a new Data Frame, consisting of one column called 'Early', with the 
# independent variable values.
new_early_wages = pd.DataFrame({'Early':[50000,55000,100000]})
# Make the prediction and print.  In other words, we are predicting what the Median
# Mid-Career Wage will be if the Median Early Career Wages are 50000, 55000, and 100000
print(lm.predict(new_early_wages))

# Plot regression line on the scatter plot using Seaborn
sns.regplot(x = 'Early',y = 'Mid', data = labor)