In this project, I try to implement a visualization of information using a physical model.

As I write this account of things, I must say this project did increase my admiration for those numerous souls who spent their time on modelling information, long before the age of computers. :-)

Data Used

I selected California (CA) as the subset of data as this area holds a consistent data collection through past years. Also, being a strong core for the technology industry population since last few decades, it forms an interesting case study for all of us.

Data for 2 pollutants, CO (Carbon Monoxide) and O3 (Ozone) has been taken into account, for the 5 years spanning 2003-2008.

I try to display the data for about 12-13 MSA (Metropolitian Statistical Area) throughout California (refer to CA MSA map) . I chose MSA data because I felt it represents a truer picture over space.

 

EPA Links for availing information used in this project:-

CA MSA map - http://www.epa.gov/ttn/naaqs/ozone/areas/maps/cacnty.gif
CA MSA data - filtered out of the 1990-2007 data sheet at http://www.epa.gov/air/airtrends/factbook.html

Materials used in the modelling

The difficult thing about this project was not only the visualisation, but that "DO NOT walk to that STORE to buy ANYTHING!" :)

Well, here is how I tried to remain true to that. A list of materials I used is shown below..

- Foam core (obtained from my sis's architectural project) and some parcel boxes
- Map pins, screws, nails, and adhesives (either stolen from my friend, or from my messy year old hardware cabinet)
- Rope caulk (the ones which we are all thankful for filling into the gaps in our homes in Chicago winter)
- String
- Some throaway blinds' components
- Paints

 

Putting it all together

A few ideas I held to while I tried to design the vis. :-

1) Allow user to get maximum information on a cursory glance.
2) Allow the more discerning user to get more information if they spend more time on this.
3) Use aspects of modelling only when necessary. Avoid additions unless they re-inforce information where needed.
4) Try NOT to make it look messy
5) Make it easy to adjust the visualisation for any given dataset (for any set of years / for any pollutant/etc.) with minimum rework

Here is a couple of quick and rough sketces I made to put down the initial concept. ( Disclaimer: I am not a good artist)

 

Elements of the design

The pollutants:

These are differentiated by using different objects for each pollutant.

I allow the user to gain information about the pollutant levels at various levels of detail:

- Lowest detail: An aggregate value over each MSA, over the representative years. This is indicated by the colour of the MSA.

- Medium detail: With a little more time, the user can visit each MSA and look at the objects on the screw ( for one pollutant), to get a breakdown of approximate trend of a pollutant. For this purpose, I use data grouped into about 4 categories, depending upon the range of data available in these years.

- Highest detail: For that ocassional user interested in exact values, I provide a scale built out of a blind. For each year, the user can easily trace out the value to the scale to obtain the value in that year.

Lowest level of detail	Medium level of detail	Highest level of detail

 

Adding to the above possibilities, overall information across California may be obtained via easy comparision:-

Using the colors, shapes on the approximate (tower) scales, and even tracing similar colored threads (for maxima/minima during each year).

 

The absolute scale on the side is varying in minimum-maximum for both pollutants. This is not exactly very clear in the visualisation, and I could not overcome that problem easily.

 

The visualisation model from different angles can be seen below.