Laboratory Assignment 4

 

Submission Date: April 11, 2016

 

In this laboratory assignment, you will implement interactive visualization of choropleth  maps. Create a choropleth map of the United States using the D3.js methodologies presented at http://bl.ocks.org/mbostock/4060606.

 

Tasks:

 

What are you supposed to do, technically speaking? The following -

 

1.     The visualization on Bostocks’ page is static.  You will need to make it “interactive”. Use sliders  and other widgets. Show trends in the form of hotspots (high employment, low employment, etc.).

2.     Note that Bostock uses employment data (available as a tsv file). Bostock uses a threshold for the unemployment trends.  Add interactors for choosing thresholds.  Further, allow two-sided thresholds for both upper and lower bounds. Here are some Javascript UI components you may find useful: crossfilter, jquery slider.

3.     Allow your tool to let the users zoom and pan across the states of the union (example).

4.     Allow your tool to let the users choose colors for display. If you think it is better to use “colors” instead of just hues as is done correctly, do change the tool. (jscolor)

5.     Allow for the use of other data from various sources. Typical other data could be temperature trends, quality of life, educational standards. Allow for the correlation of this data with employment data. You can use multiple views if you think it allows for a better visualization. Please peruse Chapter 11, 12 of the Tamara text for selection of techniques. Here is a link to a classic paper which listed the principles of creating efficient multi-view visualization systems.

 

User Evaluation:

 

Now let us list some evaluation criteria. Read the manuscript to gain a perspective on user evaluations. Paper 1, Paper 2,  Paper 3 (Section 6 gives an example of comprehensive evaluation of the tool)

 

1.     Will the tool allow user to gain an overview and gross understanding of patterns ?

2.     Will the tool allow user to gain a regional perspective of the trends ?

3.     Will the tool allow for the identification of outliers in both items and outliers ?

4.     Will the tool allow for the in-depth examination of trends between neighbors.

5.     Will the tool allow for the gross examination of correlations with other data ?

6.     Will the tool allow user to look at multivariate correlations (if it is not implemented, at least do you think that the use of “views” allow the expansion into multiple data).

7.     Will the tool allow for the generation of hypotheses ?

8.     Is the layout of the windows and widgets of the tool amenable for easy use and navigation ?

9.     Is the tool responsive enough and presents results in a manner that fosters understanding ?

10.  Do they explore multiple classification methods to find the best way to represent data? (example check table 1 in this paper )

Each team will conduct a user study. You will allow your tool to be downloaded or accessible by other teams.  You will get an evaluation as 1/0 for each of the criteria. Please use this form to evaluate the efficacy of the tool.

 

Submission:

 

Use Carmen to submit the code, a report and slide-deck (the usual two slides).  It is also required that you will make available “executable” web pages that can be executed by the instructor, the grader and other peer groups.