The Harvard Geospatial Library

1
The Harvard Geospatial Library
Optimizing Access Via ArcIMS, Portals and the
Desktop

• David Siegel
• Finding NEMO
• June 7, 2007

2
What is HGL?

• Infrastructure for cataloging, storing and
  distributing geospatial data
• A web mapping interface
• Basic cartography
• A metadata catalog
• Based on FGDC
• A standard platform for cataloging, storing and
  distributing geospatial data across Harvard
• Public access to Harvards geospatial resources
• GIS functionality on-line
• Data clipping
• Basic map rendering

3
Whats in HGL?

• Raster and Vector Data

Images from ESRI
4
Vector Data

• Locations are recorded as x, y coordinates
• Points are recorded as single coordinates
• Lines - series of ordered x, y coordinates
• Polygons - series of x, y coordinates that define
  line segments, enclosing areas.

5
Raster Data

• Rasters are composed of a grid of pixels
• Each pixel contains one value
• Scanned Maps
• Scanned air photos
• Scanned documents

1898 U.S. railroad map
6
Raster Data Continued

• Remotely sensed data
• Spatial resolution of remotely sensed imagery is
  a measure of the smallest object that can be
  resolved by the sensor or the dimension on the
  ground represented by each pixel (i.e., 30
  meters)
• Specialized cameras can capture the reflectance
  of light at several parts of the spectrum.

Landsat 7
IKONOS
7
Vector Data In HGL

• Digital Chart of the World (11,000,000)
• VMAP 1 (1250,000)
• World Vector Shoreline (1250,000)
• U.S. Census Tiger Geographies, 1995, 2000
• Digital Map Database of China (11,000,000)
• North Korea base map data (11,000,000)
• MassGIS environmental data
• City planimetric data for Boston, Cambridge,
  Newton, Somerville
• City base map data for Baghdad, Basra
• More

8
Raster Data in HGL

• USGS Historic 15 minute quadrangles for
  Massachusetts (c. 1890)
• Sotzmann maps of the northern U.S. states (c.
  1797)
• Army Map Service series L500 of China 1250,000
  scale (c. 1950)
• MODIS
• AVHRR
• More

9
All Data in HGL are Georeferenced

• A georeferenced digital image is composed of
  pixels that have geographic coordinates

1797
1835
1895
2001
City of Cambrige vector street centerline (shown
in red)
10
Data Presentation
11
Who Uses HGL?

• According to the web statistics

12
How is ArcIMS Used in HGL?

• Mapping functionality
• Rendering maps
• Obtaining the map coordinates to use in the
  metadata catalog search
• Static and dynamic map services
• Using the API to manage coordinate
  transformations
• Projecting data for map overlays on-the-fly
• Interoperability
• creating map services that can overlay with OGC
  compliant map services
• Using AXL to display data from other catalogs
• Gazetteer support
• Data clipping (vector and imagery data but not
  scanned maps)
• .

13
Data Clipping
14
HGL Software Architecture
15
Metadata Management
16
How We Provide Access to HGL

• Library Portal
• Web Interface (primary access point)
• ArcMap Desktop Tool
• Large Image Delivery Service (LIDS)
• Custom Map Services
• HOLLIS
• More

17
Library Portal
18
Web Interface - Primary Access Point
19
HOLLIS
20
(No Transcript)
21
(No Transcript)
22
Custom Map Service

• Allow custom persistent service
• Trouble with managing rendering (persistent
  rendering)

23
Linking to External Resources
24
Linking to External Resources
25
Usability Issues

• Usability issues are a constant struggle. ArcIMS
  gives you the flexibility to customize but it
  requires a lot of time and programming skills
• People expect map displays to work in familiar
  ways (AJAX)
• Usability and the back end design
• Data clipping and other GIS specific concepts are
  difficult for users to understand
• Sometimes difficult to understand the hybrid
  approach of combining mapping with textual data
  searching

26
ArcIMS Technical Issues

• Out of the box functionality may not provide a
  template that meets your needs
• Requires very specific skills to load and keep it
  running
• After about 150 map services performance suffered
• Provides OAI-PMH and Z39.50 harvesting services
  for you but requires conformance with ESRI data
  formats
• IMS software releases usually fall behind latest
  and greatest features that people want to see.
• If the functionality you want is not exposed in
  an API you have to build it yourself (Dynamic WMS
  Map Service)

27
Issues With Delivering Large Numbers and Sizes of
Data Sets

• Finding the data
• A map with 6,000 boxes showing where we have
  data is not readable
• Full text search of metadata records
• Database Management
• Upgrades (Oracle 8, 9i, 10g)
• Metadata indexing
• Data set upgrade to new schema
• Double backup
• Interfacing repositories (SDE/Oracle and the DRS)

28
Issues to Contemplate

• Open source tools
• For map display
• Metadata search
• Metadata harvesting
• Seamless data storage and retrieval
• Better tools for metadata creation
• Major points for any library to consider when
  providing digital access to maps
• Map server technology (stability, availability of
  help)
• Make sure to consider complete cycle. Do not want
  to adopt a technology that leaves you feeling
  like you bit half the worm out of an apple, Not
  practical to have 6,000 map services running all
  the time
• Out of the box geospatial library is there such
  a thing

29
David Siegel dave_siegel_at_harvard.edu Harvard
University Library, Office for Information
Systems http//hul.harvard.edu/ois/ Center for
Geographic Analysis http//gis.harvard.edu
30
Vector Data Storage in HGL
31
(No Transcript)
32
(No Transcript)