Download presentation
Presentation is loading. Please wait.
1
campaign viewer o psaap goals: o allow researchers to quickly explore campaign space o compliment and integrate into portal o visualization goals: o enable efficient visualization of multiple dimensions o offer analysis space as an interactive experience o allow to record and share discovery process o allow environment to be accessible via a browser
2
STRATEGIES FOR multivariate visualization LIMITATIONS o limited visual space o limited visual comprehension DESIGN CONSIDERATIONS o complex data better understood layered representation o reading order (levels of importance) o i.e. emphasizing/de-emphasizing different components of representation o data discretized/binned o offer >6 visual dimensions o create interactive/intuitive space
3
interactivity considerations o allow direct interaction o make objects (variables/dimensions) draggable o make rendered datapoints interactive o allow efficient interaction o easy navigation of variable-dimension coupling o via keystrokes o via clicking o via dragging
4
browser considerations o use new standards/technologies o html5 o drag/drop o canvas (vs embedded svg) o use standard client/server communication api o XMLHttpRequest o xml formatted data collections o xml formatted data o xml formatted scripts
5
canvas vs svg o canvas o + html element with direct js access o + currently faster/more stable (under chosen browsers) o - generates pixel represention o - interativity needs to be added separately o svg o - requires connections to html document (and back) o + generates vector based representation o + interactivity can be encoded in visual elements o + xml format
6
scripting considerations o allow to record interactive session o allow edit scripts o allow to save/share/publish o scripts can be stepped through o with same or different data o keep data/environment interactive along replay o make scripting/replay seamless
7
portal integration o enable collaborative data analysis o plots+scripting as part of sharable information o leading into web based collaborative dashboard o dashboard= information summary suggested actions o sample, see IBM’s Dashiki/ManyEyes wikified o data grouping o definition of batch-jobs/campaigns (custom subsets) o allows for targeted/better defined data analysis
8
application backend o goal : allow for easy access to data, optimized for grouping o object-oriented approach to group creation o strong representation of each simulation o Django implementation o Python based classes o creation of classes to represent simulations, targets, projectiles,.. o + flexibility allows for modified/new model paradigms o enables manipulation from pre-existing python code and libraries o still in development stage
9
application backend: django o implementation using jinja2 template language o increased freedom in template design o + extends django templates with custom method calls o.i.e. passing of arbitrary arguments to functions o developed with the django-command-extensions o easy database-insensitive backup o exports the database using django’s api o i.e. programmatically adds entries to the database using python o + other added features o e.g. automated way to graphically represent database backend
10
current development o developed using new html5 standards o developed in Safari o some testing on portable device (iPad) o some testing on Mozilla based browsers o implemented basic interactivity o drag/drop o keystroke navigation o script recording/playback o dynamic loading
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.