Introduction
Geospatial data can provide insight into place-based issues and trends. By mapping specific variables onto specific localities, spatial visualizations can act as important communication tools which can be easily readable by the public and by policy makers. Energy consumption, costs and environmental impacts are constant concerns. In New York City, the steam system is the largest in the world. It is considered a clean alternative energy. My map takes a look at steam consumption in Manhattan in 2010 and shows which areas, by zip code, consumed the most steam.
Inspiration
My inspirations for this visualization were pulled from existing Carto maps. The three maps that I drew inspiration from are an Oregon tsunami map, a map of oil and gas spills in Mexico and Colorado, and a map of street trees in New York City.
I was drawn to the Oregon map because it was simple, but effective. The detail used to represent changes in sea level was effective and the dots are a simple graphic that doesn’t overwhelm the viewer. The map of gas and oil spills is a dot map that acts like a density map. I thought that this was an effective technique and while the link to the live map is no longer working, the differentiation between the kind of spills, the layering, and using density of dots makes the message of this map very clear. However, the level of geographic area was a little overwhelming and detracted from the focus on the dots. Lastly, the street tree map was an example of how I might show my own New York City data. I thought that the filtering capabilities were particularly effective and that the level of granularity of the map (the illustration of city blocks) was appropriate and, at the same time, does not overwhelm the viewer.
From these maps, I envisioned that I would also make a dot map in which dot size would show the amount of steam consumed, and that my map would show clear sections of Manhattan by zip code area, but wouldn’t overwhelm the viewer with too much geographical detail.
Materials
The materials used for this lab were the data set in an excel or .csv format, a shape file of the area of interest, and Carto, a mapping application.
Methods
I found my dataset of steam consumption in Manhattan through New York City’s open data portal. Because the data was going to be visualized onto a map, I had to ensure that there was a spatial component that was organized within standard units. I chose to use zip codes, or U.S. Postal Codes. 2010 was the only year that I could find that had Manhattan zip codes associated with steam consumption, but there may be other datasets where zip code is a variable associated with steam consumption.
In order to create a map in Carto, the application requires a dataset which incorporates 3-dimensional spatial data. This is found in a shapefile which is a collection of files which contain the geometry, projection information, and attributes. These files are contained within a .zip file and must all be uploaded into Carto. Uploading the shapefile into Carto allowed for the zip codes to be represented as polygons which would show the entire area in Manhattan that fell under the specific zip code.
Once my two datasets were uploaded into Carto, Carto imposed the information onto a map. However, it showed the two datasets separately. So, I had a map of polygons which illustrated the shape of the zip code areas in Manhattan, and I had dots which indicated the location of a building which consumed steam in 2010.
In order to merge my two datasets, I had to perform an attribute analysis. The common attribute between the two datasets was the zip code variable. By creating a join between the two datasets based on the corresponding variable of zip code and by specifying my map output of zip code and steam consumption per thousand pounds of steam (Mlbs), I was able to merge the two tables and create a choropleth map which illustrated the consumption of steam per zip code. Legends were included as well as a label feature which would show the amount of steam consumed when a zip code area was clicked.
Results/Discussion
While my map is simple, I think that it also includes a lot of information that will be useful for varying levels of map reading skills. I chose a mild, but eye catching color and the variation levels between the colors are distinct. The legends allow for useful information to be highlighted and to illustrate how the outliers represented on the map translate into amount of steam consumed.
The map shows that, while New York City may have the largest steam system in the world, there are only a few areas in Manhattan which have buildings which consume the majority of the steam produced.
Future Directions
I would have liked to have added some additional features to my map, including the location of the buildings which were consuming steam and color coding the dots representing the location of the building based on whether the buildings were commercial or residential. Also, discounting areas, like Central Park, where there are no buildings would also lend clarity to the map.
My initial join of the two tables altered the data because the value type did not match. Checking to make sure that the variable and value types correspond in the join is very important. I learned that by downloading the resulting join table is a good way to check that the join was successful. My datasets were also human-readable and required minimal cleaning, but this is not always the case. The incorporation of more datasets for more complex visualizations will require that the data cleaned and organized appropriately within Carto.