Seeing the BIG picture: Quick and Dirty Data Visualisation with Ruby - Aja Hammerly @thagomizer_rb

- Motivations
  - story since we're in right after lunch. any resemblance to real people or companies is incidental
    - site slow, angry customer, angry boss, everyone panics... throw solutions at the wall.
    - WAIT
    - [Grace Hopper] "We should use COBOL" *lol*
    - go back to cubicle
    - barf numbers in boss's cube
    - ARGH
    - use pictures, you shouldn't be barfing numbers
  - Hunches aren't good enough. Data is Proof.
  - People pattern match (the more effective your communication, the more people will understand what you mean)
  - Pictures are universal (when panic happens brains shut down, visual ideas are easy to comprehend)
  - This is not a talk about Tufte beautiful diagrams.. it's about just getting things in to a visually interpretable state
- Expectations
  - Code heavy talk, thagomizer.com has the code (I'm not typing it here)
  - Practitioner focused
- Gems
  - 'graph' -- graphviz
    - Can graph associations, class layouts, etc
    - graphing dependencies is where the graph gem shine
    - based on DOT (a graphing language)
      - specifies nodes (things you're interrelating)
      - allows description of attributes of nodes
    - Basics
      - Code for the graph gem (see speaker's blog for code -- I'm not typing it here)
      - Basic DSL
        - #digraph > #edge,#node
        - exports and saving
        - Shapes
          - Triangles ( ``` node_attribs << 'triangle' ``` )
        - Color
          - ``` node_attribs << 'red'; edge_attribs << 'blue' ```
      - Pick colours with colorbrewer2.com
        - These colour schemes are built into graphviz
      - Colour Schemes
        - Categorical ... look this up on wikipedia
    - Charts
      - highcharts (free for non-commercial use, otherwise license)
        - JQuery and Highcharts.js dependencies
      - Code for generating charts with highcharts
        - See blog for code
        - You need a web page, embed the deps from googleapis
        - JQuery functions ``` $('#chart').highcharts({ ... ```
        - Going over the basic parameters to the highcharts chart specification language... see highcharts documentation for more data
        - types highcharts generates are line, spline, area, column, bar, pie, scatter, area range, area spline, column range
           - you may overlap then with multiple axes
        - Highstock is available for finanical charting
- Data Extraction
  - Formats
    - HTML/XML > Nokogiri (next talk is about Nokogiri so we won't be going into it)
    - CSV
    - Regular Expressions (rubular.com is a good tool for knowing what your RegEx is going to do)
  - Data extraction examples
    - Lots of code (see blog)
      - Extracting course descriptions and other things
      - #orphans on Graph gives you all nodes with no edges
    - When logging use sysloglogger <https://github.com/cpowell/sysloglogger> (exports PID to log message which allows one to check process balance for unicorn)
    - Highcharts does logarithmic graphs too, just specify it
- Special thanks