1 of 15

Visualization for reporting

Coulter Jones, Michelle Minkoff

​

Presentation - bit.ly/NICAR15ReporterViz

Tipsheet - bit.ly/NICAR15ReporterVizTips

2 of 15

Why chart data early in reporting?

  • Understanding your data early
    • See what’s missing
    • Discover duplicates

​

  • Find the story
    • General trends
    • Location-specific data
    • Extreme outliers

3 of 15

Tools to visualize data: Plenty of options

  • Charting
    • Spreadsheets (Excel, Libre Office, Google Docs...)
    • Online charting tools (Tableau, Silk.co, Many Eyes…)
    • Stats programs (R, SPSS, Python...)
    • Timelines, text (Timeflow, Texteture, Textalyzer…)
  • Mapping
    • Google Maps/Fusion Tables
    • Batchgeo
    • Cartodb

4 of 15

Mapping data: Quickly find the story

  • Properties owned by local rental, realtor

5 of 15

Integrity checks: Finding omitted data

  • Police depts. missing from federal data through the years.

​

Source: FBI UCR

Charting tool: R

6 of 15

7 of 15

By date: When did the story happen?

8 of 15

By date: When did the story happen?

  • Payments to doctors for a particular drug type by date.

​

Source: CMS Open Payments

Charting tool: R

9 of 15

Finding trends: Percentage change

  • Medicare reimbursement rates.

​

Source: CMS

Charting tool:

Libre Office

​

​

10 of 15

Excel visualizations

  • Homicide victimization and offending rates of juveniles and young adults, by state, 1976-99 based on the FBI's Supplementary Homicide Reports - http://www.bjs.gov/content/data/htiuss.zip
    • Repeatable bar graph in column - =REPT(“|”,A6)
    • Simple time series - Look at one state
    • Role of bars/pies
    • Stacked bar chart -for many states

11 of 15

Datawrapper - Simple viz alternative

  • https://datawrapper.de/
  • Similar types of charts to Excel, but some distinct options
  • Can be helpful if data loading issues
  • Sample data to test diff chart types right in the file
  • Can annotate notes for colleagues

12 of 15

TimeFlow

  • https://github.com/FlowingMedia/TimeFlow/wiki
  • Structure your dates’ data by thinking of what categories it belongs to
  • Navigate among different views to better understand your data
  • Project is no longer under ongoing development

13 of 15

Text visualization

  • Example: Speeches, chapters, transcripts
  • Netanyahu speech:
  • Word frequency: Textalyzer
  • Texteture: Network of how words are connected
  • DocumentCloud - entity extraction, how often words show up
    • Consider exceptions of synonyms and stop words

14 of 15

Other visualizations (non-map)

  • ManyEyes - Can explore w/data set used for Excel
  • Expands options to other viz - easier to swap among them
  • Can use for text word clouds (if you must)
  • Heat map option is helpful
  • Recently relaunched in the past yr - new interface

15 of 15

Other visualizations (map)

  • Google Fusion Tables
  • Practice mapping Toronto Community Centers
  • Change column types to be location
  • Geocoding = converting addresses to locations
  • See where trends/outliers are in your data