This should be a half hour intro talk on why people should use visualisation techniques
I NFORMATION V ISUALISATION
W HAT What is Information Visualisation (IV)? Visual encoding of abstract information to allow visual exploration /detection of patterns Can be used in tandem with statistical approaches
W HY Humans have a well-developed visual system, so take advantage of its pattern-detecting facilities Also some people just don’t trust data until “they see it with their own eyes”, or are uncomfortable with statistical measures
WHY MPG and Weight Finding patterns negatively correlated Horsepower and Weight positively correlated
D ATA S TRUCTURES Information is abstract i.e. non-physically rooted Examples include Family trees Share prices Social networks Tuple data
I NTERACTION T ECHNIQUES IV applications allow users to interact with the data, as opposed to being static screenshots (cf GraphViz) Common techniques beyond the basics include Filtering – removing, reordering and re-rendering according to selected subsets of information Linking – viewing the same data (and same filters) in different views Focusing – visual effects such as non-linear focus+context and zoom to accentuate areas of the visualisation Speed of response is vital, recommend < 50ms
I NTERACTION T ECHNIQUES Filtering works on a data set by interactively reducing the number of items that fit in the selected set. Here a house sale set of 30,000+ records is cut down to under 2,000 using the sliders on the columns.
I NTERACTION T ECHNIQUES Focusing works by giving more space to items of interest, but still retaining the ‘context’ of the unselected objects. Here the selected items in blue have increased in size.
I NTERACTION T ECHNIQUES Linking works by having data viewed simultaneously in different visualisations The linking may also apply to selections and filters Linking is closely associated with MVC architectures for separating UI and Model data. Use the same model data in multiple UI components.
W HERE Games Developers have two opportunities for using IV In the course of their work Workflow analyses Software dependencies In the game Attractive effects User attention
S OFTWARE V ISUALISATION Software visualisation – one of the first topics explored by visualisation researchers – fixing their own problems firstEick et al – SeeSoft –Developer tracker - 1992 Telea & Auber – CodeFlows SVN Visualisation - 2009 Van Ham – Call Matrices – Method Call Graphs - 2003
S OFTWARE V ISUALISATION Stand alone tools are very well, but integrating them into IDEs such as Eclipse makes them more useful (and more likely to be used)Malnati – XRay – Package CHISEL group – Creole –dependencies - 2008 Call & Method graph - 2007
L IBRARIES Developing visualisations can be time-consuming Developer Libraries Integrate common vis techniques into existing programs / websites (Prefuse, InfoVis Cyberinfrastructure) End User Libraries Drop data into visualisation (ManyEyes. Mondrian)
T HE E ND Some demos at the CISS Napier website http://www.ciss.soc.napier.ac.uk/ Q’s