Neighborhood Quality

Although I will be unable to participate in the presentation I alluded to last week, I am happy to report it proved quite feasible to prepare a package of example data in Google Earth format. I have provided a single self-contained KMZ file to my colleagues. It contains a variety of map overlays organized in folders, including labels and some descriptive annotations.

Neighborhood quality as perceived by resident students.

The interpolated surface in the screenshot above depicts the results of a survey assessing middle and high school student perceptions of neighborhood quality. My house lies just inside a hot spot. I look forward to plotting the results of the most recent edition of this survey. We’ll test the landlady’s opinion about neighborhood decline!

Posted on Tuesday, June 23rd, 2009.

Rudimentary Legends

In yesterday’s experiment with choropleth maps in Google Earth, I acknowledged the lack of classification legend. I used the Layer to KML conversion tool in that example, but today I tried the Map to KML tool. It includes certain map layout objects like legends.

Rudimentary legend generated by Map to KML conversion tool.

I don’t care for the way it looks, and it certainly needs clearer labels, but it’s a legend. The legend won’t appear in Google Earth if it was converted to graphics in ArcMap, so keep in mind that Map to KML will only preserve some elements.

Posted on Friday, June 19th, 2009.

Popular Reference

Yesterday I was thinking about how to make maps more accessible by adding reference features to the display. Today I realized that since Google Earth is such an excellent reference tool – replete with imagery, street maps, and the captivating advantage of seamless navigation – the problem of anchoring the data might be resolved by presenting it in that context. As a trial, I used the ArcToolbox Layer to KML conversion tool to save some maps in a compatible format.

Sample choropleth map viewed in Google Earth.

I think this is a useful way to lead a discussion about a place. We can present our data, yet easily and actively explore other aspects of the environment. A playful approach is engaging, and can have serious benefits: it is important to ask questions like “is this the right data?” or “what else is going on here?”. I feel this exploratory attitude is appropriate in my situation; when I present our data, I want to inspire the group to think about possible applications and variations.

The sample map pictured above is faulty in that no legend is displayed. I think it is possible to include a legend, among other things, in the KML file. I look forward to getting up to speed on Google Earth & Maps technology. I expect that in the future more Binghamton Neighborhood Project data will be made available online in this way; Google even encourages projects like ours through Outreach.

Here I advocate using Google Earth as a venue for presenting conventional choropleth maps because of the ease with which other reference layers can be explored, providing context for the data. I agree with Mark Harrower that three dimensional thematic maps (porcupine globes) are generally unsuccessful, but the debate is interesting, so check out the relevant posts by Bjørn, Leszek, and Rich.

Rich concludes with some points that reflect my own reasons for enthusiasm about this approach. In particular, he gives this example:

Draping a thematic map over topography can be a useful visualisation e.g. polygons showing erosion rate draped over a range of mountains, by eye you could then relate erosion to slope.

The sample map shown above happens to be “Percent Tree Cover” by census block. With Google Earth’s Terrain layer enabled, it soon becomes evident that the hilltops are the most heavily forested areas around the city. I am curious what other neighborhood relationships the group will uncover when we explore our demographic data!

Posted on Thursday, June 18th, 2009.

Reference Features

Next week I will help present some Binghamton Neighborhood Project data to a small committee of people interested in the state of our city’s South Side. I will be showing maps of various demographic measures collected by the BNP, addressing the question “who are the residents of the South Side?” and demonstrating the project’s ability to develop knowledge about neighborhoods.

Binghamton's South Side as viewed from Google Earth.

Today I have been thinking about how to present these maps. Spatially, the lowest common denominator for our data is the census block group. I will use census block groups as the enumeration unit for my thematic maps.

However, in this case I don’t think it is sufficient to just plot the data. Few people are familiar with census geography – you may recognize the shape of your nation or state, but do you recognize the shape of your census tract or block group?

U.S. states are easily recognized.

If the purpose of a map is to complement the viewer’s knowledge of an area, it is the map’s function – and the cartographer’s responsibility – to link the data to the viewer’s conception of the place. If the enumeration units are easily recognized – states, counties, etc. – then no further context is needed. If the data is organized at an unfamiliar resolution, additional reference features are needed to root the map in the viewer’s world.

Binghamton block groups are not so familiar, even to local residents.

Reference features can be streets, labelled landmarks, or natural features. Boundaries of more familiar regions can be informative, too, but if the region of interest is small and falls entirely within well-known borders, physical features remain the best reference. Especially pertinent are those features people encounter in their daily activities.

I think this sort of context has been absent in previous presentations of BNP data. The project has generated plenty of interest in the community, but little sustained engagement. I suspect the idea seems inspiring, but the information bewildering; it’s hard to act or base decisions on seemingly arbitrary patterns.

Our goal is not to share what we know about block groups (or any other artifact of data collection), but to share what we know about neighborhoods – places associated with homes, schools, businesses, and the routes we navigate. So, I will try to make meaningful maps by integrating our information with some of these well-known elements.

Posted on Wednesday, June 17th, 2009.