Urban Big Data Analytics

July 29, 2019

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

Pop Quiz

• Prep for 10 minutes
• Key materials we covered so far
• Principles and key elements
• Being able to explain terms

Answers to the Quiz

Assignment 4 out

• More hands-on training in R
• EDA and visualization
• Send your R code along with your plots
• Assignment 4: link

What is data visualization?

• Representing data in a graphical format
• Transforming numbers into visuals
• Good graphics are simple, intuitive, and yet tell a complex and powerful story
• Bad graphics are clusttered, embelished, and yet convey little and simple information

Typical activities in a day

The rhythm of food

Tufte's 10 Rules

• Edward Tufte - Yale professor
• Living legend and expert in data visualization
• "The da Vinci of data"
• "Chart Master"
• "The Data Artist"

1. Show your data

2. Use graphics

3. Avoid chartjunk

4. Utilize data-ink

5. Use labels

6. Utilize micro/macro

7. Separate layers

8. Use multiples

9. Utilize color

10. Understand narrative

Visualization in R - ggplot2

				# Initialize a plot
				ggplot(data, aes(x = x, y = y))
								
				

Visualization in R - ggplot2

					# Histogram	
					ggplot(data, aes(x = x, y = y) + geom_histogram()
					

					# Scatter plot
					ggplot(data, aes(x = x, y = y) + geom_point()
					

					# Adding a trend line	
					ggplot(data, aes(x = x, y = y) + geom_point() + geom_smooth()
					

					# Box plot	
					ggplot(data, aes(x = x, y = y) + geom_boxplot()	
					

Interactive mapping in R - leaflet

					# Initialize a map
					leaflet(data) %>% addTiles()
								
					

Interactive mapping in R - leaflet

					# Add markers
					leaflet(data) %>% addTiles() %>% 
						addCircleMarkers(~long, ~lat)
					

					# Visualize data
					leaflet(data) %>% addTiles() %>% 
						addCircleMarkers(~long, ~lat, radius=~value)
					

					# Add colors
					leaflet(data) %>% addTiles() %>% 
						addCircleMarkers(~long, ~lat, color=~value)
					

					# Add labels
					leaflet(data) %>% addTiles() %>% 
						addCircleMarkers(~long, ~lat, label=~label)		
					

Group Session 8

Any questions?