Visualizing Geo Data Jason Sundram, Data Scientist, PayPal - - PowerPoint PPT Presentation

visualizing geo data
SMART_READER_LITE
LIVE PREVIEW

Visualizing Geo Data Jason Sundram, Data Scientist, PayPal - - PowerPoint PPT Presentation

Visualizing Geo Data Jason Sundram, Data Scientist, PayPal @jsundram From text: To maps: Who Am I? Math Music Data! Visualization @jsundram runningwithdata.com PayPal Data Science Acquired Where.com, local recommendation


slide-1
SLIDE 1

Visualizing Geo Data

Jason Sundram, Data Scientist, PayPal @jsundram

slide-2
SLIDE 2

From text:

slide-3
SLIDE 3

To maps:

slide-4
SLIDE 4

Who Am I?

@jsundram runningwithdata.com

  • Math
  • Music
  • Data!
  • Visualization
slide-5
SLIDE 5

PayPal Data Science

  • Acquired Where.com, local recommendation & ads
  • PayPal has Petabytes
  • Types of Data
  • We’re hiring!
slide-6
SLIDE 6

What I won’t cover

Perception Map, Gina Chechak

slide-7
SLIDE 7

Three ways of looking

  • Just plot it
  • Spatial aggregation
  • Heat map
slide-8
SLIDE 8

Just Plot It

  • Benefits:
  • Easy
  • Pretty
  • Drawbacks:
  • Reference points?
slide-9
SLIDE 9

Just Plot It

Visualizing Friendships, Paul Butler

slide-10
SLIDE 10

Just Plot It

The UnFacebook World, Ian Wojtowicz

slide-11
SLIDE 11

Just Plot It

Flight Patterns, Aaron Koblin

slide-12
SLIDE 12

Just Plot It

Wispy Routes, Eric Fischer

slide-13
SLIDE 13

Just Plot It

  • Data Without Borders - NYC Data Dive
  • Jake Porway, today at 4pm
  • NY Civil Liberties Union
  • 1,193,763 stops from 2010 alone
slide-14
SLIDE 14

Just Plot It

Stop, Question, Frisk

slide-15
SLIDE 15

Just Plot It - Time

A Week of Check-Ins, Matthew Healy

slide-16
SLIDE 16

Just Plot It

  • Data:
  • Search logs - 6 million searches
  • Facebook profiles
  • Motivation:
  • Showcase demographic & behavioral targeting

Where.com

slide-17
SLIDE 17
  • All-purpose drawing
  • Java
  • Easy to get started
  • Addictive

Processing

slide-18
SLIDE 18
  • Design and edit maps
  • Slippy
  • Easy

Tile Mill

slide-19
SLIDE 19

Unfolding

  • Easily manipulate map tiles in Processing
  • CloudMade, OpenStreetMaps,TileMill

Emography, Daniel Palmer

slide-20
SLIDE 20

Putting it all together

slide-21
SLIDE 21

Three ways of looking

  • 1. Just plot it
  • 2. Spatial aggregation
  • 3. Heat map
slide-22
SLIDE 22

Spatial Aggregation

  • Choropleth = Choro (Area/Region) + Pleth (quantity)
  • Cartogram (Contiguous, Non-Contiguous, Dorling)
slide-23
SLIDE 23

Spatial Aggregation

  • Benefits:
  • Easy to see simple patterns
  • Drawbacks:
  • Perceptual issues
  • Sensitive to color/class choices
slide-24
SLIDE 24

Spatial Aggregation

Uninsured (under 65) from Stats of the Union, Fathom

slide-25
SLIDE 25

Spatial Aggregation

U.S. Agriculture, Bill Rankin

slide-26
SLIDE 26

Spatial Aggregation

United States of Craigslist, John Nelson

slide-27
SLIDE 27

Spatial Aggregation

  • Data:
  • One day of geo-tagged barcode scans
  • Motivation
  • Exploratory

RedLaser.com

slide-28
SLIDE 28

Spatial Aggregation

  • Shapefile - points, polylines, polygons
  • http://www.census.gov/geo/www/tiger/
  • pyshp - python shapefile library

California County Boundaries, geocommons.com

slide-29
SLIDE 29

Spatial Aggregation

Visualizing County Bounding Boxes rtree - spatial index

slide-30
SLIDE 30

Spatial Aggregation

data.json

slide-31
SLIDE 31

Spatial Aggregation

  • d3.js - Data Driven Documents

d3.js, Mike Bostock

slide-32
SLIDE 32

Spatial Aggregation

Poverty: The same data, the same map, different stories, ExcelCharts

emphasis equal intervals quantile average

1 2 3 4

slide-33
SLIDE 33

Spatial Aggregation

view on web Red Laser Scans per Capita

slide-34
SLIDE 34

Three ways of looking

  • 1. Just plot it
  • 2. Spatial aggregation
  • 3. Heat map
slide-35
SLIDE 35

Heat Map

  • Pros:
  • Data decides where you look
  • Better than “Just Plot It”?
  • Cons:
  • Not a lot of insight
slide-36
SLIDE 36

Heat Map

Dencity, Fathom

slide-37
SLIDE 37

Heat Map

Sightsmap, Tanel Tammet

slide-38
SLIDE 38

Heat Map

Participants by Region, OpenPaths.cc

slide-39
SLIDE 39

Heat Map

Search Volume by Language, Google

slide-40
SLIDE 40

Heat Map

  • WebGL Globe
  • Need less than 250,000 points
  • Cheat by binning (e.g. 72.12345 72.1)
  • Scale values
slide-41
SLIDE 41

Heat Map

view on web view on web Red Laser Scans

slide-42
SLIDE 42

Three ways of looking

  • 1. Just plot it
  • 2. Spatial aggregation
  • 3. Heat map
slide-43
SLIDE 43

Thank You

  • Resources:
  • http://viz.runningwithdata.com/vgd/
  • http://bit.ly/strata_vgd
  • @jsundram
slide-44
SLIDE 44

Questions?

@jsundram