Determine data sets that you want to use for your project. Begin the design of your data model to serve the questions that you are interested in pursuing. Implement this in d3, and import data that has at least  7 dimensions (total) of different types. Note: If you have not yet settled on your data for your project (and you should be thinking about it hard already), we will provide  sample datasets for you to use here. You should be able to carry out these basic operations:

1. find the maximum and minimum
2. sum values by a specified dimension (total of n in dimension d, where n is quantitative)
3. average a set of values
4. count how many records match a particular dimension criterion.
5. print out the results from the above operations , as  well as the information about the data source(s) (where they are from).
6. NOTE: as this will be web-based, these data will need to be accessible as either files on your server or as www URLs.

Many visualizations suffer significant problems that can be corrected by redesigning them on a number of different dimensions, including:

• how the data are mapped to appropriate visual features;
• visualization tasks supported;
• interaction techniques to enable those tasks and capabilities;
• appropriate data and representation models; (do the data represented allow exploration of the questions posed or support the overall goal of the visualization?)
• perceptual efficiency;
• data richness and clarity.
• etc.

For this assignment you are to choose 2 visualizations that are poorly designed. One should be designed for communication, and one should be designed for exploration. You will redesign these to address any issues you identify according to the above categories or others that may be important. You should not do this as a computational exercise but as a design exercise.

• The two visualizations you choose should deal with data issues that are sufficiently complex to require some form of analytical support (in the form of better visualization). In other words, they should not be trivially small or unidimensional data sets, but rather complex.  You'll get consideration for choosing problems that fall well into the analytics space of problem solving.
• If the visualization you choose to redesign does not originally have interaction, then do not simply propose interaction techniques to improve it. Stay within the context and redesign to the original constraints (ie, no interaction if there was none in the original).

Note:  You are not restricted to a single screen or view: in fact, you may propose a visualization of linked views, small multiples, or transitions between a set of views.

Hand In a report titled : <Yourname_a4.pdf> on Canvas that includes for each of your two visualizations:

• a description of what the visualization is meant to support
• the intended audience(s)
• the type and scale of the data
• a critique of the problems with the current visualization, grounded in what we have discussed in the course to date
• a redesign, including  images. Redesign MUST include a complete description of  interaction techniques, including storyboards of how these work as required.
• a description of the redesign, outlining all the changes that were made and justifying your choices. This should OBVIOUSLY include both a picture of the old vis and images of the new.

I should not have to say that pulling critiques and examples off websites or texts that exist will count as cheating.

Due: Sunday, Feb 11.

For this assignment you are to extend your first visualization into a  system of two or more views faceting the data according to the research questions you have identified of interest to the target audience.  These visualizations should be interactivee, and including coordinated interaction is essential (brushing and linking).

You may write this assignment using any of Java, python or Javascript/d3/jQuery, etc. If you use javascript, this must run in a browser/server configuration, so put the code on your SFU web site or another website so we can run it.If you are writing a standalone app, you must provide full running support and instructions.

You may use libraries from another source, but YOU MUST ATTRIBUTE YOUR SOURCES. I expect you to write your own code for much of this, but part of modern development is the intelligent discovery and re-use of existing resources. YOU MUST BE ABLE TO DESCRIBE WHAT THOSE RESOURCES ARE DOING, however.

Your project will require several visualizations. This is the second  visualization design and implementation. This assignment includes both visualization and a small report. Again, this will represent only a part of the visualizations you build in total for your final project, but it should be a refined and extended design towards at a useful data tool. It must:

• show multiple dimensions (1)
• use data from more than one source (1)
• user interaction, including but not limited to:
• filtering (1)
• brushing and linking (1)
• details on demand (1)
• Multiple Views (you should have multiple interactive visualizations) (2)
• Visualization Idiom (use of correct visual encoding and type to answer the research questions)   (2)
• Answer research questions: Be structured to answer at least one of the research questions you have for your project. (1)

Your submission MUST include:

• A Report: Use the description above to write the report.
• an online link to your visualization. Code and any data used.
• A design brief for your code. What classes do what ? This can be very short, but it should explain the design to us enough that if  your code does not work , we can assess enough of your work to assign you partial marks.
• a list of the external libraries/code sources. Failure to attribute will result in automatic failing grade for the assignment

  You can get bonus marks for:

  well structured and documented code (1 mark)

  Create a compressed (zip file) containing your report, code, and data to submit on canvas.