determining the effects on a neighborhood from the...

36
Determining the Effects on a Neighborhood from the Closure of a Streetcar Line Jonathan D. G. Reed Submitted to the Department of Urban Studies and Planning in partial fulfillment of the requirements for the degree of Bachelor of Science in Urban Studies and Planning at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY June 2002 Jonathan D.G. Reed, MMII. All rights reserved. The author hereby grants to MIT permission to reproduce and distribute publicly paper and electronic copies of this thesis document in whole or in part. Author ………………………………………………………..……...…………………….. Department of Urban Studies and Planning May 17, 2002 Certified by …………………………………………………………….………………….. Joseph Ferreira, Jr. Professor of Urban Planning and Operations Research Accepted by ……………………………………………………………………………… Frank Levy Chairman of the Undergraduate Committee

Upload: others

Post on 22-Jul-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

Determining the Effects on a Neighborhood fromthe Closure of a Streetcar Line

Jonathan D. G. Reed

Submitted to the Department of Urban Studies and Planning in partial fulfillment of the requirements for the degree of

Bachelor of Science in Urban Studies and Planningat the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

June 2002

Jonathan D.G. Reed, MMII. All rights reserved.

The author hereby grants to MIT permission to reproduce and distribute publicly paper and electronic copies of this thesis document

in whole or in part.

Author ………………………………………………………..……...……………………..Department of Urban Studies and Planning

May 17, 2002

Certified by …………………………………………………………….…………………..Joseph Ferreira, Jr.

Professor of Urban Planning and Operations Research

Accepted by ………………………………………………………………………………Frank Levy

Chairman of the Undergraduate Committee

Page 2: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

ii

Determining the Effects on a Neighborhood fromthe Closure of a Streetcar Line

Jonathan D. G. Reed

Submitted to the Department of Urban Studies and Planning in partial fulfillment of the requirements for the degree of

Bachelor of Science in Urban Studies and Planning

AbstractIn this thesis, I attempt to determine what effects the closing of a streetcar line has on thesurrounding neighborhood. The line in question is the portion of the MBTA Arborway(E) Line from Heath Street to Arborway (Forest Hills). Several effects of the line areexamined, including any effects the closure may have had on the number of residencesand businesses in the area, and their ratios. I also examine whether the line closure wasresponsible for any demographic and socio-economic changes in the neighborhood. It isdetermined that the line had some noted effects on the number of businesses in the area –many businesses transitioned to residences. The line closure also had a notable effect onthe habits of commuters, increasing the travel time to work, and decreasing the numberwho took public transportation to work. It had little effect on the economic anddemographic aspects of the corridor.

Thesis Supervisor: Joseph Ferreira, Jr.Title: Professor of Urban Planning and Operations Research

Page 3: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

iii

Acknowledgements

As with most things, the assistance and contributions of many people helped make thisfinal product. I would like to thank Joe Ferreira for advising me throughout the project,offering feedback on my work, and introducing me to the world of GIS. I would also liketo thank Tom Grayson for assistance in obtaining datalayers and providing tips forsoftware along the way. I would like to thank Mike Shiffer for introducing me to UrbanStudies and Planning, and for starting my interest in public transit. Finally, I would liketo thank my family and friends for all their support and assistance.

Page 4: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

iv

Contents

1 Introduction.......................................................................................................... 11.1 Background Information............................................................................... 11.2 Research Questions ....................................................................................... 11.3 Topic Statement............................................................................................. 21.4 Applicability to the Future ............................................................................ 2

2 Literature .............................................................................................................. 33 Data Sources......................................................................................................... 4

3.1 Census Data................................................................................................... 43.2 Business and Residential Data ...................................................................... 43.3 MBTA Data Layers....................................................................................... 4

3.3.1 Creation of New Datalayers ................................................................... 53.4 Other Data Layers ......................................................................................... 5

4 Methodology ........................................................................................................ 64.1 Buffering the Transit Corridors..................................................................... 64.2 Census Data................................................................................................... 6

4.2.1 Selection of Control Variable................................................................. 64.2.2 Preparation of 1990 Census Data........................................................... 64.2.3 Preparation of 1980 Census Data........................................................... 74.2.4 Business and Residence Data Input ....................................................... 7

5 Analysis................................................................................................................ 85.1 Census Data................................................................................................... 8

5.1.1 Census Data Re-buffering...................................................................... 85.2 Business and Residence Data........................................................................ 9

6 Results ................................................................................................................ 116.1 Census Data................................................................................................. 116.2 Business & Residence Data......................................................................... 15

7 Conclusion.......................................................................................................... 16A Maps .................................................................................................................. 17B Avenue Code ..................................................................................................... 24

B.1 fc.Calc ......................................................................................................... 24B.2 fcWaitDlg ................................................................................................... 25B.3 fc.WaitDlg.Open......................................................................................... 25B.4 Other Scripts............................................................................................... 26

C Pictures .............................................................................................................. 27

Page 5: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

v

List of Figures

Figure 1. Percentage of People Using Public Transportation as Means ofTransportation to Work............................................................................... 11

Figure 2. Percentage of People with a 20 minutes or less Commute to Work .. 12Figure 3. Median Household Income ................................................................ 13Figure 4. Percentage of Vacant Housing Units ................................................. 14Figure 5. Population Density of Boston (1990) Showing E Line and C Line

Buffers ........................................................................................................ 17Figure 6. Map of Jamaica Plain, showing the block numbers and street names

along the former E Line Right of Way....................................................... 18Figure 7. Land Use Changes from 1984 to 1996............................................... 19Figure 8. Land Use Changes from 1984 to 1987............................................... 20Figure 9. Land Use Changes from 1987 to 1991............................................... 21Figure 10. Land Use Changes from 1991 to 1994............................................. 22Figure 11. Land Use Changes from 1994 to 1996............................................. 23Figure 12. fcWaitDlg......................................................................................... 25Figure 13. A PCC Car formerly used on the E Line........................................... 27Figure 14. The Existing Tracks Along South Street........................................... 28Figure 15. Coolidge Corner Looking Outbound................................................. 29Figure 16. Coolidge Corner Looking Inbound ................................................... 30Figure 17. The Centre Street Business District ................................................. 31

Page 6: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

1

1 Introduction

The availability of mass transit is important for an urban environment.

1.1 Background InformationThe first step in the closure of the Arborway Line occurred in 1959, when theMetropolitan Transit Authority (MTA) opened the Riverside Line. This new line was thelongest of all the branches of the streetcar lines. To keep up with the demand for serviceon the new line, the MTA had to remove cars from other lines, which resulted in carshortages throughout the entire system. In 1963, both the Watertown and Arborway lineswere targeted for closure to alleviate car shortages, however this problem wastemporarily alleviated the following year. In 1964, the MTA became the MassachusettsBay Transportation Authority (MBTA), and was able to receive much more funding dueto the larger number of cities and towns served by the system, and new sources of federalfunding. The streetcar lines, or "Green Line" as it was now called, still faced carshortages. As many cities had scrapped their fleets of PCC (Presidents ConferenceCommittee) streetcars (Figure 13), there was no source of vehicles which the MBTAcould use to alleviate the shortage.

In 1969, the situation was declared an emergency, and the “A” line to Watertown wasclosed on a temporary basis, which later became a de facto permanent closure. In 1973,the MBTA ordered its first Light-Rail Vehicles (LRVs), in the hopes that they could beused elsewhere on the system and thus allow the PCC cars to be used solely on theArborway Line. Throughout the 1970s, service was sporadic at times and the line wasclosed several times. In 1980, the line was closed for two years in order to rebuild themedian strip between Northeastern and Brigham Circle. In 1982, the line was re-openedagain, however the PCC cars were in a terrible state of repair, despite a federal grant in1973 to rehabilitate them. In late 1985, the MBTA announced that the Arborway Linewould be closed “temporarily” to allow re-paving of South Huntington Avenue and toallow the rebuilding of the subway incline at the Huntington Avenue portal which, in itsprevious state, could not withstand the weight of the new LRVs. The line was slated tore-open in 1987, but it did not. Instead, it was replaced by the Number 39 bus, whichruns from Forest Hills to Back Bay Station. The tracks, however, still remain in the streetalong the former line. (Figure 14)

In the past year or so, the MBTA has shown an active interest in restoring service to thecorridor, and has completed several studies to determine what methods will be used. Therestoration of streetcar service, favored by most of the population, is one option - anotheris CNG (Compressed Natural Gas) bus service, with larger waiting areas, similar to thenew Silver Line.

1.2 Research Questions

Did the closure of the Arborway Line have an effect on housing occupancy, housingvalues, or household income? Did the closure of the line have an effect on businesses

Page 7: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

2

along the corridor? Did the closure of the line and its replacement bus service have aneffect on the utilization of public transit along the corridor?

1.3 Topic Statement

It is my hypothesis that the closure of the Arborway portion of the MBTA “E” Lineresulted in a decrease in the number of people using public transportation to go to work,and an increase in travel time to work for people living along the corridor. I also believethat the closure of the transit line would lead to an increase in the number of vacanthousing units in the area, and a decrease in the median household income. Lastly, theclosure of the line would result in a decrease in the number of businesses along theformer line, possibly accompanied by a shift of businesses towards the northern end ofthe corridor where the line still remains.

1.4 Applicability to the Future

There a numerous benefits that would result from having an answer to these questions.The question of whether bus service is a suitable replacement for light rail is one thatcould prove useful for the MBTA should they decide to reduce service on one of theother light rail lines. It could also prove useful to other transit agencies throughout thecountry that are considering suspending light rail service in favor of bus service.Knowing the socio-economic effects from the closure of the line could also determinewhat effects might result if streetcar service were restored to Jamaica Plain.

Page 8: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

3

2 Literature

Much has been written on the subject of public transportation and its effects onneighborhoods. Transit has often been a defining force behind the development of newurban environments. This is particularly true with light rail, which gave us the term“streetcar suburb”, referring to new, mainly residential, areas of the city that sprung upalong or at the ends of streetcar lines. To quote Robert Carson, “Builders assumed thatwith the appearance of transit service in undeveloped areas residential building wouldfollow.”1

There are numerous pieces of literature available which cite the benefits of light rail,including the fact that it can navigate short-radius curves, can run along streets in privateor shared right-of-way, and is significantly cheaper than rapid transit.

Little attention has been given, however, to the effects that result from the closure of atransit line. Surely, if a “… well-designed street running Light Rail system …. Cangreatly add to the quality of city life”2, if such a system were removed, the neighborhoodaround it should suffer. The benefits provided by a transit line – such as faster access toother areas of the city, an increase in development of the area – should undoubtedlyreverse themselves if the line were to disappear.

Using census data and a time series of business and residence data, I will attempt to provethat there are significant effects on a neighborhood from the closure of a transit line.

1 Carson, Robert B. What Ever Happened To The Trolley? University Press of America: Washington,D.C., 1978.2 Barry, Michael. Through The Cities: The Revolution in Light Rail. Frankfort Press: Dublin, 1991.

Page 9: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

4

3 Data Sources

Several data sources were needed to perform this analysis. Census Data from 1980 and1990 was needed. I also needed to find a source that would allow me to extract the countand locations of businesses and residences along the former transit corridor. Finally,since the Arborway section of the “E” Line had been closed since 1985, it did not exist onany MBTA transit layers. Therefore, the layer would need to be created.

3.1 Census Data

The 1990 Census data was made available by the Computer Resource Lab (CRL) in theDepartment of Urban Studies and Planning (DUSP). The Summary Tape File (STF) 3Ain dBase IV (DBF) format was provided by the CRL, as was a shape file of 1990 CensusBlock Groups.

The 1980 Census Data was extracted from the CensusCD 1980 by GeoLytics,Inc.(GeoLytics, Inc. P.O. Box 10, East Brunswick, NJ, 08816. 2000}

For each of the two years, I obtained data for Suffolk County and Norfolk Country. Thelatter was required because much of the “C” Line runs through the City of Brookline,which is located in Norfolk Country.

3.2 Business and Residential Data

The Cole City Directory (a/k/a Blue Book Cross Reference) contains listings ofresidences and business ordered by street name. Using directories from previous years, itwas possible to determine the numbers of residences and businesses along the corridor.As the entries also had street addresses associated with them, it was also possible togeocode this information, as discussed in Chapter 4.

My original plan was two use data from 1984, 1985, 1986, and 1987. However, thiswould not provide enough time for businesses and residences to react to the closure of theline and relocate. Therefore, the current data set is from 1984, 1987, 1991, 1994, and1996.

3.3 MBTA Data Layers

Both MassGIS and the CRL provided data layers with the current MBTA lines and stopson them, however these data layers did not completely align with the block group andstreet centerline files obtained from the Census Bureau. Due to this discrepancy, thefollowing workaround was employed.

Page 10: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

5

3.3.1 Creation of New Datalayers

For the purpose of this project, I was concerned primarily with two sections of the MBTAlight rail system. The first was the section from Arborway (Forest Hills) to Heath Streetin Jamaica Plain. The second, which was used as a control variable for analysis of theCensus Data, is the section of the “C” Line on Beacon Street from Park Drive toCleveland Circle. Both of these sections of transit line follow street centerlines. Itherefore decided to extract portions of the street centerlines from the TIGER files, anduse those to designate the two transit corridors. To do this, I obtained street centerlinefiles (extracted from TIGER data) from MassGIS. I then selected the appropriateportions of the streets using MBTA maps, and Boston Street Maps. Locations of transitlines were also double-checked against USGS aerial photographs.

3.4 Other Data Layers

One necessary, yet unavailable datalayer was the boundaries of Jamaica Plain. AsJamaica Plain is a neighborhood of the City of Boston, and not an autonomous politicalentity, finding such a layer proved difficult. An ad hoc layer was created, using thegenerally agreed-upon boundaries of Jamaica Plain. These boundaries are WashingtonStreet and Columbus Avenue to the East, Brookline/Norfolk County to the East,Allandale Street and the Arborway to the South, and Boylston and Tremont Streets to theNorth. A higher degree of accuracy is not necessary, as the border merely serves as abasis for a grid to create surfaces (see Section 5)

Page 11: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

6

4 Methodology

4.1 Buffering the Transit Corridors

The selection of the buffer size for census data analysis proved to be more difficult thananticipated. In general, ¼ of a mile is considered the greatest distance a person will walkto a transit line. Therefore, I selected ¼ mile buffers around each line. However, whenselecting the block groups, that caused problems, which are addressed in Section 5,Analysis.

4.2 Census Data

In order to eliminate the possibility of any regional or local trends (such as a prosperousreal estate market) contaminating the final results, all Census Data analysis wascompared against a control variable. The control variable in this case was the “C” branchof the MBTA Green Line.

4.2.1 Selection of Control Variable

For the “control” transit corridor to be effective, it would have to be a light-rail line atgrade. This eliminated the Red, Orange, and Blue lines, as well as the elevated andsubway portions of the Green Line. The Mattapan High-Speed Line was a potentialcandidate, as it used the same PCC cars that ran on the Arborway Line until the latterclosed in 1985. However, the Mattapan Line runs through very economically depressedareas, and is mainly used as a shuttle between Mattapan Square and Ashmont station,with few people embarking or disembarking at the intermediate stops. The “D” branch ofthe Green Line was eliminated because it runs in a private, partially depressed right-of-way and has limited stops. This left the “C” and “B” Lines. The “B” Line runs downCommonwealth Avenue, which is much wider than either Centre Street or Beacon Street.It also has a much narrower median, and the narrow platforms, and high fencesdiscourage people from waiting for the train. The “C” Line made the ideal choice for acontrol line, as it has a large, pedestrian friendly commercial district – Coolidge Corner –which is comparable to the Centre Street area on the Arborway section of the “E” Line.The Coolidge Corner area can be seen in figures 15 and 16, and the Centre Streetbusiness district appears in Figure 17.

4.2.2 Preparation of 1990 Census Data

The 1990 Census Data needed to be prepared in order to conduct analysis using it. Eachtable in the STF3 was located in a separate DBF file. These DBF files could be joined tothe block group layer by means of a common Federal Information Processing Standard(FIPS) code unique to each record. However, the block group layer contained a FIPScode string that was a concatenation of the state, county, tract, and block group codes.

Page 12: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

7

This concatenated code did not exist in any of the STF3 tables. I created an Avenuescript (shown in Appendix B) to scan the STF 3 tables, and concatenate the relevantcodes in the appropriate manner to create the unique identifier that would allow the tableto be joined to the attribute table of the blockgroup layer. The Avenue script waspackaged into an extension called FIPSCalc for ease of use.

4.2.3 Preparation of 1980 Census Data

The 1980 Census Data, stored on the CensusCD, needed to be extracted for use inArcView. The CensusCD provided shapefiles of the 1980 block groups. These blockgroups were derived from the 1992 TIGER/Line files, and so contain some sliverpolygons. This is the best digital coverage of the 1980 block group boundaries, however,so the sliver polygons cannot be avoided. The extracted shapefiles were in DecimalDegrees, and therefore needed to be projected into the 1983 State Plane, Massachusettscoordinate system, using meters as the units. This ensured compatibility with the 1990Census Datalayers and the other layers used.

The 1980 Census Data also did not use the STF3 codes from the census, so some researchwas needed to determine what the equivalent STF3 codes were for the 1980 field headers.

4.2.4 Business and Residence Data Input

The Cole City Directory does not exist in electronic form for the years with which I amconcerned. The microfilm versions were obtained from the Boston Public Library, andpaper copies of the relevant pages were made. From there, the street address, zip code,and a field indicating whether the record was a residence or business were entered into aspreadsheet. The spreadsheet was converted to a DBF file for use with ArcView. Usingthe MassGIS provided TIGER files containing street centerline and geocodinginformation, each address was geocoded. There were no failed matches, and only oneinteractive match was required. The rest of the addresses matched with perfect scores onthe first round.

Page 13: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

8

5 Analysis

5.1 Census Data

I compared 4 measurements for the Census Data Analysis: Vacant Housing, MedianHousehold Income, Travel Time To Work, and Means of Transportation to Work.

The Median Household Income from the 1980 Census was adjusted using the ConsumerPrice Index (CPI). The 1980 Census uses data collected in 1979, and the 1990 Censususes data collected in 1989. The CPI between 1989 and 1979 is 1.70798, so the medianhousehold income values from 1979 were multiplied by 1.70798 to yield adjusted valuesfor 1989 dollars. There is still a large discrepancy between the 1979 and 1989 values forthe entire Boston Metropolitan Area (Norfolk and Suffolk Counties). The averagehousehold income in 1979 for the area was $24,463, compared with $31,803 in 1989.This is most likely the result of a national trend.

Public Transportation values also required adjustment. The values for 1990 includedmeans of transportation such as Ferryboat that were not included in 1980.

5.1.1 Census Data Re-buffering

After making an initial block selection, I compared the blocks selected across both years.The shapefile for the 1980 block group boundaries contained many “sliver polygons”.These anomalies result when one data layer does not match up with another data layer.The primary reason for this was the introduction of the Topologically IntegratedGeographic Encoding and Referencing (TIGER) System in the 1990 Census. This wasthe result of an effort by the census bureau to improve geometrical accuracy for Censusdata. Due to this large increase in the accuracy of measurement from 1980 to 1990,several discrepancies occurred between the two layers. The 1990 block group layer wascreated in 1995 by MassGIS directly from the 1990 TIGER/Line files. The 1980 layer,was created from the TIGER/Line files released in 1992. This is the only set ofTIGER/Line files that attempts to correlate the 1980 and 1990 boundaries and identifythese “sliver polygons” created when census blocks changed and measurementsimproved.

Selecting block groups which intersected the ¼ mile buffer proved to be troublesome,since several block groups were shared between both lines. Selecting block groupswhose centers fell within the ¼ mile buffer was also problematic, as the sliver polygonsin the 1980 boundary layer caused ArcView to incorrectly determine the center of somepolygons, and to select polygons which fell far outside the buffer.

After further investigation, I increased the buffer size to ½ mile, and altered the selectioncriteria to only include those blocks which fell completely within the new buffer. Thatallowed for a much better block selection, and solved the problem of the sliver polygons.

Page 14: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

9

5.2 Business and Residence Data

For each of the 5 years, I added a field to the table. This field was an integer that was 1 ifthe point was a residence, and 10 if it was a business. I then created a grid theme fromthe Jamaica Plain border theme. This grid contained 5-meter grid cells. I then selectedthe point theme for 1984, and interpolated a grid based on the new field I had added. Iused the default methods and settings for the interpolation. This resulted in a surface thathad lighter colors (values closer to 1) in areas of high residential concentration, anddarker colors in areas of high business concentration.

Interpolation of the point data required that I make some decisions. For the business andresidence data, I am primarily concerned with effects along the immediate transitcorridor. Along the line, there are few businesses on the side streets, and I only had datafor the businesses and residences that fronted onto the corridor. I chose a buffer of 50feet (radius) along the corridor, which would allow for the width of the street, and thebuildings on either side of it. That buffer was then also converted to a grid theme.

The grid was interpolated with the buffer surface as the analysis mask. This process wasrepeated for each year of data, for a total of 5 surfaces. I then used the Map Calculatorfeature of ArcView to subtract the 1996 layer from the 1984 layer. The resultingcalculation provided a layer that had negative numbers (as high as -9) where businesseshad given way to residences, and positive numbers (as high as 9) where residences hadgiven way to businesses. The output from Map Calculator was re-classified by StandardDeviation.

Several assumptions were made when interpolating the points. First, the method usedwas Inverse Distance Weighted (IDW). This assigns a lower weight to points that arefurther away from the point being evaluated. Second, I chose to evaluate the 12 nearestneighbors, as opposed to points that fell within a certain radius. The reason for this isthat the analysis was restricted to a very narrow corridor (namely, the width of the streetand adjacent buildings), and the radius would necessarily have to be less than that. Thatwould result in very skewed results in areas where there is a large distance between oneaddress and the next. Initially, I thought that locations where more than one pointoccupied the same location (i.e.: multi-family residences, or office buildings) would posea problem. However, after reviewing descriptions of the methods used in performing theanalysis, I determined that ArcView used the 12 nearest points, even if those points werein the exact same location.

This still poses a problem for three areas along the corridor. The Boston HousingAuthority’s (BHA) South Street Apartments at 125 South Street and Farnsworth House at90 South Street both contain large numbers of residences, however each also contains onebusiness (namely, the management office). This however, will not pose a problem in thefinal results, since those business remained across the entire time period in question. Thesecond location is at Lakeville and Beaufort Terraces, on Centre Street, and the third isthe VA Hospital at Heath Street. Both of those buildings have the same problem as the

Page 15: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

10

BHA developments, however any errors introduced by this can also be disregarded, asthe business points remained across the time period.

Page 16: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

11

6 Results

6.1 Census Data

After using ArcView’s “Select By Theme” feature to select all the block groups that werecompletely within the ½ mile buffers I had created, I calculated average values for eachline within each year. Calculating the average across each block group would yieldinaccurate values, as the population across block groups is not homogeneous. Therefore,the percentage per block group was weighted based on the population of that blockgroup. This weighting was performed by multiplying the percentage by the population atthe block group level, and then dividing by the total population of the selected blockgroups.

The percentage of people taking public transportation to work along the E Line corridordeclined from 36.223 in 1980 to 35.289 in 1990. Along the C Line, it increased slightlyfrom 35.263 to 35.823. The sample size is something which must be taken intoconsideration. A sample size of 20,000 is required to obtain a less than 1 percent marginof error. The universe of the count (in this case, working population), exceeds 20,000, soa 1% change, although small, can be counted as statistically significant. The generaltrend throughout the metropolitan area was a slight decrease between 1980 and 1990 inthe number of people taking public transportation to work.

Figure 1. Percentage of People Using Public Transportation as Means ofTransportation to Work(N.B.: The graph has been enlarged to better illustrate the difference. Note the range of the Y-axis)

Means of Transportation To Work

34.634.8

3535.235.435.635.8

3636.236.4

1980 1990

Year

Per

cen

tag

e o

f P

eop

le

E Line

C Line

Page 17: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

12

Travel time to work also exhibited a change. In 1980, the percentage of people along theE Line who had a 20 minute or less commute to work increased from 33.978 to 35.686 in1990. Along the C Line, it increased from 31.908 in 1980 to 34.652 in 1990. These arealso weighted percentages using the method described earlier. The general trendthroughout the metropolitan area was a slight increase in the percentage of people with a20 minute or less commute to work in 1990 as compared with 1980.

Figure 2. Percentage of People with a 20 minutes or less Commute to Work(N.B.: The graph has been enlarged to better illustrate the difference. Note the range of the Y-axis)

Travel Time To Work

30

31

32

33

34

35

36

1980 1990

Year

Per

cen

tag

e o

f P

eop

le

E Line

C Line

Page 18: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

13

Median Household Income did not exhibit a significant amount of change along the ELine (as compared with the C Line) between 1980 and 1990. Along the E Line, the 1980income (in 1990 dollars) was 23,786, whereas the 1990 income was 33,164. Along the CLine, the 1980 income (in 1990 dollars) was 26,672, whereas the 1990 income was34,538. As mentioned in Section 5.1, however, the large discrepancy between the twoyears for each line is the result of a larger trend.

Figure 3. Median Household Income

Median Household Income

05000

100001500020000

25000300003500040000

1980 1990

Year

Do

llars E Line

C Line

Page 19: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

14

The final data I looked at was housing unit vacancy. The percentage of vacant housingunits, however, did not change much on one corridor as compared to the other. In 1980,on the E Line, 6.716 percent of the housing units were vacant, up to 7.079 in 1990. In1980, on the C Line, 4.75 of the housing units were vacant, up to 5.036 in 1990.

Figure 4. Percentage of Vacant Housing Units

Vacant Housing Units

01234

5678

1980 1990

Year

Per

cent

age

of V

acan

t H

ou

sin

g U

nits

E Line

C Line

Page 20: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

15

6.2 Business & Residence Data

For the purposes of this section, refer to Figure 6, which shows Jamaica Plain and the ELine corridor, along with street names and block numbers.

I made several comparisons of the results from analyzing the business and residence datacreated from Map Calculator. I chose to only consider values as significant if they were2 standard deviations or greater in either direction. I first considered the overall changefrom 1984 to 1996. This map (Figure 7) showed that a significant amount of changeoccurred in the 300-400 block of South Huntington Avenue. In that area, many locationsthat were businesses in 1984 were residences in 1996. Similar patterns were present inthe 400-500 block of Centre Street. Along the rest of Centre Street and South Street,there were some isolated pockets of change, however they are not significant enough tocount, given the time period involved. One small change, however, occurred in the 200block of South Street. That block shows that a large cluster of residences in 1984 werebusinesses in 1996.

I also examined the transition between each of the 5 samples, to determine exactly whensome of the changes exhibited in the overview analysis might have occurred. From 1984to 1987 (Figure 8), two main changes are noticeable. In the 100 block of South Street,many businesses have become residences. In the 200 block of South Huntington Avenue,the same phenomenon has occurred, however at the very end of the South HuntingtonAvenue section of the corridor, a large number of residences have become businesses.

From 1987 to 1991 (Figure 9), no significant changes occurred. There are isolatedpockets, but not enough to form a conclusion.

From 1991-1994 (Figure 10), many areas changed from businesses to residences. Theseinclude the 100 block of South Street, the 500 block of Centre Street, and 300 block ofSouth Huntington Avenue. This could be attributed solely to the economic slump around1990. However, it is also likely that the closure of the transit line had a significant effecton the patronage of businesses, yet they were able to remain open for short period oftime. The economic slump could have merely hastened the inevitable.

From 1994-1996 (Figure 11), there are some conversions from residences to businesses.These include 530-550 Centre Street, and 10-40 South Street. This could be the oppositeof what occurred during 1991 to 1994 – namely that businesses were taking advantage ofthe slightly better economic climate and were re-opening. Regardless, the number ofbusinesses that changed to residences over the 12 year period far outweighed the numberof changes from residence to business. Theoretically, large changes such as these couldbe caused by catastrophic events (such as destruction of a block by fire), but no suchevents occurred during the 12 year period in question.

From the overall analysis, it would appear that much of the changes from businesses toresidences occurred between 1991 and 1994, 5-10 years after the transit line closed.

Page 21: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

16

7 Conclusion

The closure of the transit line had a small, yet significant effect on the use of publictransportation as a means of transport to work. At the outset, this may seem like anobvious conclusion. However, when one takes into account that the census definition of“public transportation” includes all major modes such as heavy rail, rapid transit, lightrail/streetcar, and bus, it provides us with a very interesting conclusion. From theanalysis of the census data, it would appear that the replacement bus service along theformer corridor (which follows the former streetcar line exactly) was not adequate, atleast from a commuter point of view. The marked increase in travel time to work alsoimplies that the streetcar line provided faster transportation than the replacement busservice, or any other substitute for that matter.

The closure of the transit line had little discernable effect on the demographics of theneighborhood. While efficient public transportation is an important part of an urbanenvironment, is not critical for survival. A lack of a one-seat ride to the rest of Boston isnot enough to discourage people from moving into a neighborhood, nor is it enough toencourage current occupants to relocate elsewhere. The closure of the line certainly didnot change the economic demographics of the neighborhood. My original thought wasthat the termination of one-seat ride service to downtown Boston (or anywhere else onthe MBTA Rapid Transit Lines) would limit the number of accessible jobs, and thuslower the median household income. This turned out not to be the case. While the lossof the line may have made it inconvenient for residents to travel to some jobs, it did notmake it impossible. Further study in this area could include examining housing values,rents, and housing sales data to determine if any of these factors were affected by theclosure of the line.

The closure of the transit appears to have had a significant effect on the land use alongthe corridor. Many businesses converted to residences during that time. Further study isneeded, however, to determine whether these effects were more a result of the transit lineclosure or the economic environment at the time.

There are many opportunities for further study in the area. A study of nearby Roxbury,and the eastern portion of Jamaica Plain, that attempts to determine any effects from theclosure of the elevated (which resulted in the Orange Line moving to the east of itsformer location) would be useful. That, combined with data from my study, might helpto paint a more accurate portrayal of the results from the closure of the E line. It mightalso help to determine whether the observed changes in commuter behavior was dueentirely to the closure of the E line, or whether it was also due in part to the closure of theEl. A more detailed study on transportation effects that builds on the results here mightalso prove to increase understanding of the effects from the closure of a streetcar line.

Page 22: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

17

A Maps

Figure 5. Population Density of Boston (1990) Showing E Line and C Line Buffers

Page 23: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

18

Figure 6. Map of Jamaica Plain, showing the block numbers and street names alongthe former E Line Right of Way

Page 24: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

19

Figure 7. Land Use Changes from 1984 to 1996

Page 25: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

20

Figure 8. Land Use Changes from 1984 to 1987

Page 26: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

21

Figure 9. Land Use Changes from 1987 to 1991

Page 27: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

22

Figure 10. Land Use Changes from 1991 to 1994

Page 28: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

23

Figure 11. Land Use Changes from 1994 to 1996

Page 29: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

24

B Avenue CodeI created an extension in order to calculate the FIPS codes for the 1990 STF3A tables, sothat they could be joined to the block group coverage. The components of that extensionappear below.

B.1 fc.Calc

'Script to Calculate the FIPS codes from 1990 Census tract files'Copyright 2001 Jonathan Reed [email protected]

theTable = av.GetActiveDocif (theTable.GetGUI <> "Table") then msgBox.Error("No active Table!"+NL+"Please make a Table active and try again.", "") exitendtheVtab = TheTable.GetVTab

if (theVtab.CanEdit) then result = MsgBox.YesNo("This will add the field [Stcntrbg2] to this table. Proceed?","FIPSCalc", TRUE) if (not result) then exit endelse msgBox.Error("This table is not editable!"+NL+"Check file permissions.","") Exitend

hasStrcntrbg2 = theVTab.FindField("Stcntrbg2")if (hasStrcntrbg2 <> nil) then msgbox.Error("This table already has a field [Stcntrbg2]!"+NL+"Cannot proceed.","") exitend

hasStatefp = theVtab.FindField("Statefp")hasCnty = theVtab.FindField("Cnty")hasTractbna = theVtab.FindField("Tractbna")hasBlckgr = theVtab.FindField("Blckgr")

if ((hasStatefp = nil) or (hasCnty = nil) or (hasTractbna = nil) or (hasBlckgr = nil))then msgbox.Error("This table is missing the necessary fields."+NL+"Refer to documentationfor more information.","") exitend

geowait=av.finddialog("fcWaitDlg")

geowait.FindByName("fcWaitDlgStatTxt").SetLabel("Searching...")

geowait.open

theBitmap = theVTab.GetSelection

SearchStr = "[Tractbna].Count > 4"'Add the FieldtheField = Field.Make("Stcntrbg2", #FIELD_CHAR, 12, 0)theVtab.setEditable(TRUE)theVtab.AddFields({theField})'Look for Tractbna entries more than 4 charstheVTab.Query(SearchStr, theBitmap, #VTAB_SELTYPE_NEW)

'theVTab.UpdateSelection

'Calculategeowait.FindByName("fcWaitDlgStatTxt").SetLabel("Calculating...")geowait.FindByName("fcWaitDlgStatTxt").Update

Page 30: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

25

theVtab.Calculate("[Statefp] + [Cnty] + [Tractbna] + [Blckgr]", theField)'Flip SelectiontheVtab.GetSelection.Not

'theVtab.UpdateSelection

'Calculate Others, adding "00"

CalcString = "[Statefp] + [Cnty] + [Tractbna] +" + "00".Quote + "+ [Blckgr]"theVtab.Calculate(CalcString, theField)

'Save editsgeowait.FindByName("fcWaitDlgStatTxt").SetLabel("Saving...")geowait.FindByName("fcWaitDlgStatTxt").UpdatetheVtab.setEditable(FALSE)theVtab.GetSelection.ClearAlltheVtab.UpdateSelectiongeowait.close

B.2 fcWaitDlgFcWaitDlg is a dialog box, since the script was packaged as an extension. The textversion of the Dialog Code is long, so for the sake of brevity, I will include only an imageof the dialog box.

Figure 12. fcWaitDlg

B.3 fc.WaitDlg.Openthedialog=self

AVUpperLeft = av.ReturnOriginAVCenter = avUpperLeft + (av.ReturnExtent / (2@2))

halfDialogWidthHeight = theDialog.ReturnExtent.ReturnSize / (2@2)

MovePoint = AVCenter - halfDialogWidthHeighttheDialog.MoveTo(MovePoint.GetX, MovePoint.GetY)

ifn = "$AVHOME/tools/bitmaps/wait.bmp".AsFileNameif (File.Exists(ifn)) then if (Icon.isFileValid(ifn)) then III = Icon.Make(ifn) theDialog.FindByName("fcWaitDlgIconBox").SetIcon(III) endend

Page 31: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

26

B.4 Other ScriptsThe rest of the extension consists of the variations on the standard Avenue scripts for usewith extensions, namely “Make”, “Install”, and “Uninstall”.

Page 32: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

27

C Pictures

Figure 13. A PCC Car formerly used on the E Line

Page 33: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

28

Figure 14. The Existing Tracks Along South Street

Page 34: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

29

Figure 15. Coolidge Corner Looking Outbound

Page 35: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

30

Figure 16. Coolidge Corner Looking Inbound

Page 36: Determining the Effects on a Neighborhood from the …web.mit.edu/uis/theses/jdreed/jdreed_final_02may.pdfneighborhoods. Transit has often been a defining force behind the development

31

Figure 17. The Centre Street Business District