Talks & Papers‎ > ‎

NICAR 2011 ...........shortcut to this page: http://goo.gl/z4khx

To present,
  1. open the links below in separate tabs.
  2. Start up Google Refine and move it to the appropriate tab

Use Shift-Command-F to enter full-screen
Use Option-Command-RightArrow and -LeftArrow to move between tabs while in full-screen

50 minute live demo / hands-on workshop


Introduction

Context of Structured Data at Google

  • Google has made it's reputation organizing unstructured data. Now we're turning our attention to helping people get more out of structured data.
    • [unemployment rate ireland]
  • Google Public Data & Explorer is a great example. Data from the World Bank, UN, and other trusted sources is digested to create richer search results. Hans Rosling and his son Ola created the Gapminder visualization, and Ola brought his passion for public statistics to Google to create the next generation of the tool. Now you can upload data to use it too.
  • Spreadsheets in Google Docs of course is a great cloud-based tool for working with structured data, and even publishing it like the UK Guardian did here for the Drug War Deaths
  • Google Refine is a power tool for working with messy data
  • Google Fusion Tables is an easy way to make charts and maps. A modern data management web app making it easy to host, manage, collaborate on, visualize, and publish data tables online

Example of Fusion Tables

  • When the UK Guardian DataBlog wrote about the deaths in the Mexican Drug War, they offered a map. The distribution of incidents on the map makes it quite clear that this is effecting the entire country.  They used Fusion Tables' map tools to turn a bunch of data into thousands of points on the map.
    • click link to see map fullscreen
    • click to see the data in Fusion Tables
    • click Visualize > Map

Overview of features

  • [feature list] Now I'm going to demonstrate a lot of great features in Fusion Tables.
  • all skill levels - this talk is focused on what you can do just in the UI

Demo: Timelines and Pie charts

Pie Chart

How much variety has there been in popular female baby names? 
    1. Filter: Rank = 1
    2. Aggregate: Female Name
    3. Visualize: Pie
Not much variety in the top female names; Mary holds her own.

Timeline

How popular has "John" been over time? 
    1. Filter: Male name = John
    2. Visualize: Timeline
John was incredibly popular for the first part of the century, and has rapidly fallen out of favor since. 

Also: 
  1. Filter: Male name contains Greg  ->Gregory

Demo: Mapping lots of points


Demo: Custom intensity map

Where do boundary files come from?


Top Examples of Fusion Tables

Using Fusion Tables web UI only

Getting into Google Maps API & more


Demo: Mapping Restaurant inspections in NYC

  • NYC Data Mine offers the data
  • bring WebExtract.csv into Refine, load only 1000 rows
    • Create Location column:
      • join([cells.BUILDING.value, cells.STREET.value, cells.ZIPCODE.value], " ")
    • export as .csv
  • bring into Fusion Tables  - example
    • add attribution data
    • sort Desc on Inspection Date
    • Visualize > Map , wait for geocoding
    • apply style by buckets:
      • 27 & less is green, 28 and up is red
    • customize infowindow:
          <b>{DBA}</b><br>
          {Location}<br>
          Score: {SCORE}
    • Make Unlisted
    • Get Embeddable Code --> fullscreen map URL  
    • Bonus:
      • merge in Cuisine.txt for Cuisine Code
      •  add to infowindow


Demo: Intensity map of licensed parking facility spots per NYC zip codes

Comments