Urban Big Data Analytics

July 30, 2019

Instructor: Andy Hong, PhD
Lead Urban Health Scientist
The George Institute for Global Health
University of Oxford

Assignment 4

• Assignment 4: link
• Any issues? errors?
• Very similar to our group session yesterday
• Due tomorrow (Weds) by 5:00pm
• Send your R code as well

What is statistical learning?

• A set of tools for understanding data
• Based on probability and statistics
• Supervised learning: using inputs to estimate outputs
• Unsupervised learning: inputs but no outputs
• Good for explanation
• Linear regression, logistic regreassion, classification

What about machine learning?

• Algorithm-based models
• Supervised and unsupervised learnings
• Good for prediction
• Random forest, support vector machine (SVM), gradient boosting

Statistical Learning vs Machine Learning

Simple regression

Regression Line

Regression explained

Ordinary Least Squares

Multiple regression

Multiple regression

Interactive regression plot

Simple Regression Demo

					# Simple regression
					m1 = lm(data = gapminder, lifeExp ~ income)
					summary(m1)
					
					# Call:
					#   lm(formula = lifeExp ~ income, data = gapminder)
					# 
					# Residuals:
					#   Min      1Q  Median      3Q     Max 
					# -18.032  -3.948   1.314   4.217   9.300 
					# 
					# Coefficients:
					#   Estimate  Std. Error t value            Pr(>|t|)    
					# (Intercept) 67.70194625  0.54603278  123.99 <0.0000000000000002 ***
					#   income       0.00025963  0.00002118   12.26 <0.0000000000000002 ***
					#   ---
					#   Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
					# 
					# Residual standard error: 5.522 on 185 degrees of freedom
					# Multiple R-squared:  0.4482,	Adjusted R-squared:  0.4452 
					# F-statistic: 150.3 on 1 and 185 DF,  p-value: < 0.00000000000000022
					
					

Multiple Regression demo

				# Multiple regression
				m2 = lm(data = gapminder, lifeExp ~ income + population)
				summary(m2)
				
				# Call:
				#   lm(formula = lifeExp ~ income + population, data = gapminder)
				# 
				# Residuals:
				#   Min      1Q  Median      3Q     Max 
				# -17.939  -3.903   1.410   4.129   9.379 
				# 
				# Coefficients:
				#   Estimate      Std. Error t value            Pr(>|t|)    
				# (Intercept) 67.602268543445  0.560501264567 120.610 <0.0000000000000002 ***
				#   income       0.000260247911  0.000021214915  12.267 <0.0000000000000002 ***
				#   population   0.000000002244  0.000000002797   0.802               0.423    
				# ---
				#   Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
				# 
				# Residual standard error: 5.528 on 184 degrees of freedom
				# Multiple R-squared:  0.4501,	Adjusted R-squared:  0.4441 
				# F-statistic: 75.31 on 2 and 184 DF,  p-value: < 0.00000000000000022
				
				
				

Group Project Part 3

Any questions?