Reverse Engineering New Urbanism
This tutorial was concieved in collaboration with Professor Brent Ryan
Development dimensional standards provide a normative model for the form of places. The SmartCode is a new approach to developing such a normative model that embodies the principles of a branch of place-based scholarship known as New Urbanism. The intention of this tutorial is take some of the dimensional standards of the SmartCode and see how well they relate to specific places. This will give us new way to understand a place in terms of how well it represents new urbanist concepts.
The basic procedure demonstrated here is an application of our approach to Place Based Modeling for Scholarship and Decision Support.
We have defined our intention and our source of scholarship in the introductory paragraph. What is left to do includes: Being specific about the concepts we are going to investigate; and to look for measurements and observations (data) that represent these. Then we will organize these data into a schema that will enable us to represent what has been measured about the place in terms that are comparable with those used in the smartcode. This schema will allow us to generate maps of our place that will provide the basis for comparison of the formal ideal with an actual configuration. Furthermore, we will be able to create a modified version of our schema to experiment with calibrating the smartcode to better reflect the place, or to rezone the place to better reflect the code. In either case, our model will allow us to evaluate our work in a systematic way.
By working through this demonstration we will see that the simple conepts we have pulled out of the Smartcode are far from a comprehnsive embodyment of what SmarrtCode intends. Furthermore, we will see that the measurements and observations that we have access to are not perfect reflections of the smartcode concepts that we have chosen to explore. Therefore, after doing our best to develop new information, before assigning any level of confidence to it, we must consider whether our data and procedures fit the formal concepts we intend to model either from the SmartCode, or on the ground. Whether we find that this model is valid or not, we will certainly have learned something about the needs for information and procedures that would be required for doing a better job with this.
This tutorial assumes a familiarity with the concepts and procedures discussed in the preceding tutorials:
- Place Based Modeling for Decision Support
- GIS Data and Metadata
- Beginning a GIS Database
- Elements of Cartographic Style
- Representing a Place and its Context
- Mapping with Nominal Class Data
- Mapping with Quantitative Data
To explore a model that evaluates the consequenses of altering dimensional
regulations in terms of buildout development capacity, see
A model of Urban Development Capacity.
One of the fundamental devices used in the SmartCode is the
SmartCode Transect, which
expresses a typology of urban places that relate to eachother as a spatial
progression. The siz types of places described in the SmartCode are
distiguished and blend from one to the other in complex ways, that are
explained in the SmartCode which is over 200 pages long. For the purpose
of this exercise, Professor Ryan has chosen just two concepts to diferentiate
the districts: Lot Coverage and Unit Density.
One of the fundamental devices used in the SmartCode is the SmartCode Transect, which expresses a typology of urban places that relate to eachother as a spatial progression. The siz types of places described in the SmartCode are distiguished and blend from one to the other in complex ways, that are explained in the SmartCode which is over 200 pages long. For the purpose of this exercise, Professor Ryan has chosen just two concepts to diferentiate the districts: Lot Coverage and Unit Density.
|SmartCode T Zones||Unit density||Lot coverage|
|T2 rural||< 1.5 units/ac||< 55%|
|T3 sub-urban||1.5 < 3 u/ac||55 < 65 %|
|T4 general urban||3 < 5 u/ac||65 < 75 %|
|T5 urban center||5 < 9 u/ac||75 < 85 %|
|T6 urban core||> 9 u/ac||< 85 %|
The idea of unit density in the SmatrCode has a lot to do with the number of parking spaces required. In the case of housing, a unit is considered to be a household, and in the case of commercial land, a unit corresponds with 1000 square feet of gross floor area. Lot coverage is an idea that is related to the amount of space between buildings, which has a lot to do with pedestrian accessibility, and the general feel of a place.
Creating the Schema
We have a lot of data from the town of Concord and the City of Somerville. Unfortunately neither town has given us explicit measurements of number of units or percent of lot coverage. We will either have to give up at this point, or figure out how to make inferences about these conepts based on information that has been given to us.
The first thing we will do is create a new geodatabase and make copies of the layers that we want to experiment with. Some of the information we will be using comes frm the parcels layer, some comes frm the buildings layer. In order to find the lot coverage, we have to union these two layers of geometry together and pass information about the area of the building footprints to the apropriate parcel records. This is a process that is easier to to show in a geoprocessing model than it is to explain in text. So we will add the model to our map, and explore it. Geoprocessing models are a means of saving sequenses of operations. A great thing about a model that is designed well, is that we will see, once we have the model running for areas in Somerville, the same model will also work in Concord!
Initialize the Schema
- Set your Tools>Geoprocessing Options as described in the Geoprocessing Cheat Sheet It is not necessary to do the steps under the heading Practicing...
- Add the toolbox, pbc New Urbanist Tools to your tool box window. You will find this toolbox in the folder gis/tools in your sample dataset.
- Right click on model number 1 and choose Edit Double-click on the yellow Create Geodatabase box inside and take a look at what it is doing.
- Run this model to create a new geodatabse in your scratch folder.
- Now look at Model number 2. This model figures out the amount of area of each parcel that is covered by buildings. It also adds new fields to hold our estimate of the number of units per each parcel, and the estimated unit density, and a new field where we can assign parcels to SmartCode T Zones.
- Zoom in to an area of the map that you want to investigate, and set the Extent parameter for the model to be Same as Display
- Run the entire model number 2.
- Take a look at the data that was produced.
Critique the estimation of Lot Coverage
Now we have generated soem new information that we may use to represent the smartcode concept of Lot Coverage. We must take responsibility for this. It certainly is not perfect. Is it good enough? For some tips on how to evaluate the utility of data such as this, see Critique of Data and Metadata.
Estimate the Number of Units in Each Parcel
Now the free ride is over. You have to decide how to model the Smart Code concept of Unit Density. There are a number of ways you might approch this: as a simple product of the Gross Floor Area, or you may compute the units based on a lookup of the parcel land use code in the department of revenue land use class lookup table (which you will find in the project. Whatever method you choose you should end up by calculating a new value for the Est_Units (estimated units) column of your parcel table. From this you can calculate unit density for each parcel. Hint: there are 43,560 square feet in an acre.
As usual, you should evaluate the new information you have created. Your calculation of units and unit density is certainly not perfect. Do you think that it errs toward estimating too high or too low? Do you think that the inevitable errors make a material difference in terms of the interpretation you would make based on these data?
Extra Credit: how could you check your estimation of units using the census data included with this project? If you want to see this question really taken to extermes, see this GIS Study of Parcel Versus Census Estimations of Population Density by Corey Zehngbot (March '10).
Make Maps that Portray how the parcels may be divided into T Zones
Make maps that portray the parcels based on their lot coverage and unit density. Use the classes that are recommended by the Smart Code. Please consult Elements of Cartographic Style and associated material on making thematic maps. Interpret these maps as indicated in Professor Ryan's instructions.
Create a Proposed Zoning Scheme
Your assignment (handed out on a separate sheet has some requirements for creating a new zoning map. YOu can achieve this by selecting groups of parcels and then calculating their New_T_Zone attribute using the field calculator.