Harvard Returns to Geography

1
Harvard Returns to Geography

• The reasons for creating and
• the prospects for the
• Center for Geographic Analysis

2
Harvard and Geography

• Harvard University closed its Geography
  Department in 1948. Since then --
• Quantitative geography has gained much ground in
  the 1950s (http//www.infoplease.com/ce6/world/A08
  58357.html)
• Geographic Information Systems started in the
  1960s when computers became available
  (http//envstudies.brown.edu/Thesis/2001/james/gis
  history.html)
• Geoinformatics, geospatial analysis, geotech
  boomed in the past two decades as fruits from the
  crossbreed of geography and information technology

3
Geographic Analysis in Harvard

• Geographic analysis continues to occur in
  Harvard
• The Harvard Map Library and Harvard Geospatial
  Library
• Datasets in the social sciences from HMDC
• Urban planning and landscape from the GSD
• Public health data from HSPH and the Ellison
  Institute
• Environmental sciences from the University Center
  for the Environment
• The China Historical GIS project

4
The Harvard Map Library

• Catalog of geographically referenced data
• Repository for storing data
• Infrastructure for distributing data

Andover historical map
5
Harvard Geospatial Library
A catalog and repository of geospatial data
within the Harvard library system
6
Datasets in the Social Sciences from Harvard MIT
Data Center
7
Urban Planning and Landscape from the Graduate
School of Design
8
Urban Planning and Landscape from the Graduate
School of Design
9
Data from Harvard School of Public Health and the
Ellison Institute
Contamination Plume
Predicted particulate concentration
10
Environmental Sciences from the University Center
for the Environment
Coal fueled power plants
11
The China Historical GIS Project
Tax quota in the year 1077
12
Why?

• Because it is estimated that 80 of all data has
  a spatial component a component that has been
  largely neglected

13
The Creation of theCenter for Geographic Analysis

• Many faculty and staff across the university have
  joined discussions during the past two years
  concerning
• improving access to spatial data,
• support for research employing geospatial
  analysis,
• and curriculum development.
• With support and encouragement from the Provost,
  the FAS Dean and the GSD Dean, the Center for
  Geographic Analysis has been established in the
  new Institute for Quantitative Social Science.
  (http//hgl.harvard.edu8080/HGL/html/CGA_announce
  ment1.pdf http//www.news.harvard.edu/gazette/200
  5/10.20/26-bol.html)

14
CGA Goals

• The Center for Geographic Analysis at Harvard
  University is a newly established program
  focusing on research and education in the field
  of spatial analysis and geographic information.
• The Centers goal is to work with entities across
  the university to
• strengthen university-wide geographic information
  systems' infrastructure and services
• provide the common platform for the integration
  of data from diverse sources and knowledge from
  multiple disciplines and
• enable scholarly research that would use, improve
  or study geospatial analysis techniques.

15
A technology platform that supports research and
teaching in all the fields that employ geospatial
analysis across the university
16
CGA Staff Responsibilities

• Supporting research and teaching that relies on
  geographic analysis
• Helping to select and administer Harvard-wide GIS
  software site licenses
• Collecting and disseminating spatial datasets
  currently residing in isolated locations around
  the university
• Working with the Harvard MIT Data Center to
  create links between the Harvard Geospatial
  Library and the Virtual Data Center project
• Maintain the Harvard GIS website
  (http//www.gis.harvard.edu)

17
CGA Activity Plan

• Organize a guest speakers series to promote
  academic exchange in geographic analysis
• Create positions for post-doctoral fellows who
  will work with research projects across the
  university
• Create internships for undergraduates and
  fellowships for graduate students who apply
  spatial analysis in their research
• Provide teaching facilities for geographic
  analysis through the computer labs in the Center

18
Faculty Position and Curriculum

• In creating the Center the University has also
  committed itself to raise funds for senior
  faculty positions that will give us greater
  strength in such fields as geospatial analysis,
  geoinformatics, and geography.
• We are concerned with adding to the undergraduate
  and graduate curriculum across the University and
  look forward to working with current faculty in
  DEAS, FAS, GSD, and HSPH to help develop and
  support the appropriate courses, course modules,
  and labs.

19
CGA Governance

• The Center has faculty steering committee of
• Peter Bol, FAS, Director of the CGA
• Gary King, FAS, Director of the IQSS
• Niall Kirkwood, GSD, Chair of Landscape
  Architecture
• Carl Steinitz, GSD
• Peter Rogers, DEAS
• Louise Ryan, HSPH
• Robert Sampson, FAS and
• Dan Schrag, FAS, Director of the Harvard
  University Center for the Environment.
• There is also a professional technical staff
  committee composed of professionals from across
  the university.

20
(No Transcript)
21
CGA as a Hub

• There are already a number of research groups and
  centers, departments, and schools which employ
  geographic analysis, some of which already have
  professional staff with expertise in GIS. In
  addition the Harvard College Librarys Map
  Collection has GIS specialists.
• The Center for Geographic Analysis serves as a
  hub that supports and coordinates with these
  various spokes by providing technical and
  design expertise and by increasing the
  capabilities of the Harvard Geospatial Library as
  a common repository for geospatial datasets.

22
Support Research and Teaching

• There is a substantial list of projects across
  the University that seek technical support.
• It is expected that when research projects
  require ongoing staff support that grant
  proposals will include this in the budget.
• The Centers staff has already begun to provide
  support in formulating grant proposals and
  conducting analysis for on-going research
  projects
• Thailand local economic development
• Germany labor productivity
• Amazon forest fragmentation
• We also support teaching and provide technical
  consultation to students.

23
The Effects of Protected Forest Areas on Local
Economic Development in Villages of Chiang Mai
Province, Thailand

• By Kate Eman, Kennedy School of Government
• This project analyzes how protected area policies
  have affected economic development in the context
  of villages in Chiang Mai Province in Northern
  Thailand. 
• It compares the growth of selected household
  assets and employment rates for villages inside
  and outside different types of designated forest
  protection areas over a period from 1986-2003,
  using bi-annual survey data from the Thai
  Community Development Department. 
• Geospatial data plays an important role in
  choosing villages in the comparison or "control"
  group that are similar in terms of geographic
  characteristics (including elevation, slope, soil
  type, and proximity to major water bodies) to
  those villages that are inside of protected areas.

24
Thailand Project Tasks

• Collect and calibrate data
• Conduct geospatial analyses
• Make maps, and
• Provide technical tutorial

25
Thailand Project Analysis
The project calculates five geographic data
variables for roughly 1,500 villages in the
Chiang Mai province.

• Average Elevation within 1.5 km radius
• Average Slope within 1.5 km radius
• Distance to closest major river or perennial
  stream
• Distance to border of nearest conservation area
• Soil type composition within 1.5 km radius

26
Thailand Project Mapping

• Visualize major steps of analyses
• Make thematic maps for publication

27
Thailand Project Result

• Export result into database for further
  statistical analyses

28
Thailand Project Tutorial Documentation

• Help researchers to have a better understanding
  of the methodology
• Make it easier for researchers or graduate
  students to repeat the procedure if they want to
  change variables or just want to practice GIS
  skills
• Document major steps that are useful to similar
  projects

29
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis
By Nicola Fuchs-Schündeln, Assistant Professor of
Economics and Rima Izem, Assistant Professor of
Statistics

• The study analyzes the reasons for the stubbornly
  low labor productivity in former East Germany and
  aims to distinguish between two main causes
  worker characteristics (e.g. skills) vs. job
  characteristics (e.g. capital or infrastructure).
• The study uses a spatial labor market model that
  allows for commuting. The discontinuity of
  unemployment rates in proximity to the former
  border will determine which factor is more
  influential in the unemployment rate.

30
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis

• The study uses county level data and tools of
  spatial econometrics to empirically analyze the
  slope of the unemployment rate along the former
  border, and calibrate the model to match the
  observed slope.
• Euclidean (e.g. as the crow flies) distance
  measurements distort proximity to labor markets
  because commuters rarely travel through the air.
  A drive-time matrix provides a more accurate
  measure of labor market spatial relationships.

31
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis

• Task 1 Create Network Dataset
• The source of the data were layers of Germanys
  Autobahns and National Roads from MACON GFK
• Each road segment in the source layers was
  attributed with distance. The estimated travel
  time on each segment was calculated using speed
  assumptions for the different road classes.
• A Network Dataset was built using ArcGIS Network
  Analyst. The dataset had to be rebuilt after it
  was discovered that 15 of the junctions in the
  road network were dropped due to offsets in
  vertices. The Integrate tool was used to correct
  the offsets.

32
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis
Task 1 Create Network Dataset
Original
Integrated
Autobahns and National Roads
33
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis

• Task 2 Create Layer of District Points
• The source of the data were layers of Germanys
  Districts (Kreis) and Municipalities (Gemeinden)
  from MACON GFK. Population data was attached to
  the Municipalities layer.
• The research uses data from 439 Districts.
  Network analysis requires origin and destination
  points. Therefore, a representative point had to
  be created for each district.
• The centroid of each district was rejected as a
  representative point due to its lack of relevance
  to settlement patterns. The most populous
  municipality in each district was queried and its
  centroid was deemed to be the best representative
  point, due to the smaller size of the
  municipalities.

34
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis
Task 2 Create Layer of District Points
Sample District with Municipalities Shaded by
Population
35
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis

• Task 3 Create an Origin-Destination Cost Matrix
• The Network Dataset and the representative points
  layer were used as inputs for an OD Cost Matrix
  analysis in Network Analyst.
• The resulting table of nearly 200,000 routes was
  parsed to create one record for each unique pair
  of points. The analysis was performed twice
  once with time as the impedance and once with
  distance as the impedance. The data was exported
  to ASCII files for use in statistical analysis
  software.

36
What Causes the Low Labor Productivity in East
Germany? A Spatial Analysis
Task 3 Create an Origin-Destination Cost Matrix
37
Amazonia Forest Fragmentation Impact on
Precipitation

• By Paul Moorcroft, Associate Professor of Biology
• This project compares landuse/landcover
  classification maps in different years to
  determine land use change patterns.
• It models forest fragmentation,
• and its impact on precipitation.
• The forest edge effect is related to
• forest fire risks, which in turn
• affects the hydrologic cycle.

38
Amazonia Forest Fragmentation Impact on
Precipitation Tasks

• Image classification for land cover
• from LandSat images, delineate non-vegetated land
  (water, bare soil, pavement) non-forested land
  (farmland, meadows) and forestland
• Spatial analyses
• land cover change detection
• land cover fragmentation
• distance to edge calculation
• forest degradation quantification
• Simulation results visualization
• Mapping
• 3-D animation

39
Amazonia Forest Fragmentation Impact on
Precipitation Data
Red band 7, mid-infrared Green band 4
(near-infrared) Blue band 2 (visible green)
LandSat Image, summer 1990
40
Amazonia Forest Fragmentation Impact on
Precipitation Data
Red band 7, mid-infrared Green band 4
(near-infrared) Blue band 2 (visible green)
LandSat Image, summer 2000
41
GoogleEarth File Development and Training for
Chinese History Students

• An ArcObjects procedure was developed to create
  GoogleEarth place files (KML files) from the
  China Historical GIS database.
• Freeware KML generation tools were not sufficient
  because they produced ASCII files, thereby
  corrupting the Chinese characters stored in the
  database.
• The procedure created a text file using the UTF-8
  character set to maintain the Chinese characters.
• A Chinese History class was given a one-hour
  introduction to GoogleEarth using the sample
  files.
• The class included instruction on placemark
  creation and description editing.
• The class was in preparation for coursework in
  which students will identify market towns using
  GoogleEarth.

42
GoogleEarth File Development and Training for
Chinese History Students
43
Technical Support for GIS Data Development and
Analysis
An undergraduate student needed assistance in
calculating the distance from census tracts to
state borders. His research involved border
effects for unemployment and welfare benefits.
After an hour of instruction and guidance
regarding ArcToolbox features and ArcGIS
techniques, he was able to complete the analysis
independently.
44
Technical Support for GIS Data Development and
Analysis
Later in the week, the student requested
assistance in eliminating ocean and international
boundaries from the state boundaries layer. With
minimal prompting, he learned basic ArcEditor
techniques and created the layer independently.
45
Center for Geographic Analysis
1737 Cambridge Street, N319 Cambridge, MA
02138 617-496-6102 (voice) 616-496-5149
(fax) wguan_at_cga.harvard.edu

• Contact
• Wendy Guan, Ph.D.
• Director of GIS Research Services
• Center for Geographic Analysis
• Harvard University