Eugen Taso - Tufts University



Eugen Taso

Assignment #6

1. The goal of your analysis

In this assignment, I analyzed the public transportation network (buses and subway stations/stops) relative to residential parcels in the City of Somerville. I wanted to see how many residential parcels are in the city, where they’re located and, of those, how many are located within 0.25 miles (1300 feet) of a bus stop and 0.5 miles (2500 feet) of a subway stop. This analysis would be useful not only for a public transportation project for Somerville (trying to see how many people from residential areas have easy access to public transport), but also for determining housing values, LEED certification points opportunities, etc. For this assignment I loaded parcel data (2008) from Somerville from the M drive into ArcMap. I also loaded the MBTA data for bus stations and subway lines and stations from MassGIS (on the M drive) and then added additional layers to make the map readable (town boundary, EOT roads clipped to Somerville, adjacent towns) and used the resulting map for my analysis.

Please note that I was going to try to do a mini-preview of my final project (nuclear sites in New England and the impact assessment of possible nuclear disasters on population, infrastructure and environments within a 5-mile, 25-mile and 50-mile radius of each power plant). However, given time constraints and personal circumstances, I decided to stick to easier, readily available layers for this assignment and settled on the Somerville transportation project.

2. The steps you went through (i.e., the queries you performed, in order)

a. First, I loaded the parcel data for Somerville, alongside the bus stops and the subway stops in or adjacent to Somerville. I also added the road outlines from EOT roads and the adjacent town for geographical orientation. The results are below:

[pic]

Figure 1 – Somerville Parcels layer, bus stops, subway stops, EOT roads and adjacent towns

b. I took the parcels layer, chose Select from the toolbar menu and created a layer from the original parcels layer for residential parcels only, using the selection tool:

[pic]

Figure 2 – Selecting by attribute (residential parcels)

[pic]

Figure 3 – Selecting by attribute – Residential Parcels Layer

c. I took the newly created residential layer and selected by location in order to get a layer of residential parcels with proximity to the bus stops layer and the subway stops layer. I took 0.25 miles distance from bus stop and 0.5 miles from a subway stop, in accordance to LEED criteria (which is why I see this as a useful project for looking at LEED certification). The results are shown below (I obtained two new layers, one for bus stops and one for subway stops, and used transparency to show the overlap):

[pic]

Figure 4 – Selecting by location (0.25 miles from bus stop, 0.5 miles from Subway stop)

[pic]

Figure 5 – Residential parcels within 0.25 miles from bus stops

[pic]

Figure 6 – Selecting by location (0.5 miles from Subway stop)

[pic]

Figure 7 – Residential parcels within 0.5 miles from subway stops

[pic]

Figure 8 – Overlaid residential parcels within 0.25 miles from bus stops and 0.5 miles from subway stops

d. Next, I looked at the statistics for several selected features. I first looked at the residential layer as a whole to assess the values of the properties (I chose to look at sale price), and then I dug deeper and isolated the residential parcels in the bus stop layer (0.25 miles away from a bus stop), as shown below:

[pic]

Figure 9 – Statistics for selected features (sale price of property for residential parcels)

[pic]

Figure10 – Statistics for selected features (sale price of property for residential parcels close to bus stops)

e. I then summarized my data by an attribute field in order to see how many of each category are contained in the data, and that the average value is for the specific categories. I looked at price of property and at use designation, as shown below:

[pic]

Figure 11 – Summary by an attribute field value (sale price of property for residential parcels)

[pic]

Figure 12 – Summary by an attribute field value (land value of property for residential parcels)

f. For the final task, I added a field to the attribute table for the residential layer in order to be able to perform a field calculator function. I wanted to look at the price per square foot, so I added a new field to the attribute table. Then I calculated the price divided by the square foot and presented the results for residential parcels below:

[pic]

Figure 13 – Adding a new field (price/square foot)

[pic]

Figure 13 – Using the field calculator to determine the value/sq foot

[pic]

Figure 14 – Resulting new field and values after field calculator

3. A map or maps that show the results along with summary table(s) you create

[pic]

Figure 15 – Somerville residential parcels proximity to mass transit

4. A note on why the results of your analysis are or might be incorrect

There are several reasons why the results of this analysis may be incorrect. The first and most obvious is human error. It may be possible that during the analysis I selected the wrong attributes, loaded the wrong layers or summarized the incorrect fields.

The second caveat is the accuracy of the data. The analysis is only as good and as complete as the data it is based on. Are all the parcels correctly defined in the Somerville GIS dataset? Are we missing any values? Does that affect our price/land value analysis? Does that impact our map? Could someone base buying a property based on this analysis, given the potential incompleteness of the dataset? In this case, there are several specific question marks that could be raised about the accuracy of the data, and therefore about the accuracy of the analysis performed. Namely, in Figure 11, we can see that we have over 1000 parcels with a price of 0, which, given that they are residential only, is strange. Similarly, we notice in Figure 14 that the top field (and other, if we were to scroll down or sort ascending) are 0. this is of course based on the price 0 divided by the area, but again, it is hard to believe that I could just walk to those residential properties in Somerville and get them for free.

The third is that price may not be influenced necessarily by proximity to T stops or bus stations, but rather by location close to Davis/Porter squares (grocery stores, public safety, etc), schools, common spaces, open spaces, etc. Therefore, the analysis in Figure 9 and Figure 10, summarizing the average prices for residential and residential within 0.25 miles of a bus stop, should be taken with a grain of salt.

Finally, it is worth noting that this analysis may be more useful if further refined. As it stands it is a very general exercise meant to allow for honing skills and exercise commands learned in class. It may not necessarily be providing major insight into the land values associated with public transport proximity for residential parcels in the city. However, looking at a specific area of the city and comparing the parcels that are within 0.25 miles of the bus stop to ones that are slightly further out (notice on the map that most if not all residential properties are within a walking distance from the transit stops) to see whether that impacts the price may be of further interest. In addition, the properties analyzed are almost all within the 0.25 – 0.5 mile radius of the T-stop, so the price difference is likely to be explained by other factors.

Also, please note that the additional special join attribute (optional) was not applicable to this exercise, and therefore was not conducted.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download