Who am I?

Managing director of a small Swiss company:
Klokan Technologies GmbH.
Innovative geo applications in culture heritage.
Software development services & customization
of open-source projects, participation on grants.
Clients: BL.uk, NLS.uk, David Rumsey (USA),
NationaalArchief.nl, MZK.cz, Swiss libraries, ...
Petr Pridal <petr.pridal@klokantech.com>




www.oldmapsonline.org

  National Library of Scotland, Edinburgh, UK, December 13th 2012
www.oldmapsonline.org




The easy-to-use gateway to
 historical maps in libraries
     around the world.
Search engine for maps

Library catalogs & fulltext search fail on maps.
Search by geographic location / area on a
modern world map
Indexing maps from many different library
collections
One click to view the map in high resolution
Free to access and open for contributions
Contributor friendly
Bring online visitors to the websites of
participating institutions to view the maps
Promote the map libraries and their content
Do proper crediting: logo and links back
Store only minimal metadata & thumbnails
Boost the number of online visits of maps
Make it easier to discover the scanned maps
Partners
More to come...

Open for new contributors. Institutions can
let index their map collections. It is for
free. The contributor retain all the rights.
Ongoing negotiation with additional
libraries.
Submission of maps from individuals at a
later point (technically more problematic)
Contribute
High resolution (300dpi+ or 2 Megapixel+)
maps with a zoomable viewer online
Permanent URL with direct access - no
password or payment required
Metadata with DublinCore elements
supplied in a spreadsheet (XLS). Later OAI-
PMH harvesting.
Bounding box (MARC 034, DCMI BOX, ...)
http://project.oldmapsonline.org/contribute
Bounding box
Bounding box encoding
                MARC: field 034

034: $$dE0055700$$eE0102900$$fN0474800$$gN0454900
255: $$cE 5°57'00"-E 10°29'00"/N 47°48'00"-N 45°49'00"



            DublinCore: DCMI Box

DCMI BOX: westlimit=5.9559; southlimit=45.818;
eastlimit=10.4921; northlimit=47.8084



            ISO 19115 / FGDC / OGC
How to create such records?


       BoundingBox Tool
                 + Geoparser Tool

       Georeferencer Tool
BoundingBox Tool

BoundingBox Tool: Cataloguing while
holding the physical maps - manual
selection of the area on an online map
(used by Swiss libraries, David Rumsey, ...)
Geoparser: Semi-automatic rapid process -
with suggestion of the area by analyzing
existing textual metadata.
http://boundingbox.klokantech.com
Georeferencer Tool
Enrichment of the online scanned maps
A tool which runs in a web browser
An online service - no software installation
Uses volunteers for metadata enrichment
Successful pilots in several institutions
16
17
Competitive
Crowdsourcing
Georeferencer
It allows to turn the existing online scans into
geodata in a web browser
Metadata enrichment, with 3D visualization as
a side product
Support true collaborative editing
It retains log of all operations (a la Wikipedia)
- you can adjust it with a few clicks
Library can receive GeoTIFF / WMS / ...
Project launched in the end of February in
New York and London
Over 30.000 visitors per day on the website
after launch
Two servers in a load balancer to handle the
traffic after launch.
An amazing feedback from general public
Publicity
Social Networks: Twitter, Facebook
Google Developer Showcase + Google I/O 2012
BBC History magazine online slideshow
La Stampa (Milan, Italy) newspaper
Gizmodo, Ars Technica, The Verge
Wired.jp (Japan), About.com Genealogy,
Playtech.r (Romania), Portsmouth News
(UK), ... TV spot in Catalonya: link
120.000
        high-resolution historical maps
           covering the whole world


(by the end of the project it is going to be close to 150.000 probably)
During the last 10 months
   we have helped over

       400.000
people to find maps and visit
   online map libraries
We have encouraged
  all libraries holding maps
to think about their metadata
  in a geographical context.
Technology
Based on the MapRank Search technology
http://www.mapranksearch.com/
Released in 2010 by Klokan Technologies
for Swiss project Kartenportal.ch, improved
for Czech libraries and for David Rumsey
Useful for large metadata catalogs +
geospatial digital repositories.
Details in the article published in D-Lib:
doi:10.1045/september2011-oehrli
Technology
MapRank ranking: an advanced algorithm
calculating coverage similarity of geographical
area of query and area covered by every
document and combining it with map scale,
time range, predefined categories and
traditional full text search.
Technology
Default user interface developed with Closure
JavaScript Library + Closure Compiler.
Server component: FastCGI, C++, runs on UNIX/
Linux
Data indexing implemented specifically to
provide ranked real-time results even for large
databases. Traditional R-Tree indexing or
Lucene are failing.
with # of records:
                                                          70.000

      Performance                                         125.000
                                                          250.000
                                                          500.000
           for 70.000 test queries                        1.000.000

900



675




                                                              Time in seconds
450



225



  0                                  old implementation
      new implementation
Google indexing
1-3 position on Google for “old maps” phrase
The intuitive geographical search interface uses
AJAX, it is only one page for Google bots :-(
The project must expose the metadata in a way
indexable by traditional full-text search engine
Phrases such as the “Old maps of Edinburgh”,
“Old maps of British Isles”, “of ...”, ...
World coverage, high quality geodata required
Linked network of geodata




Natural Earth Data - possible export as RDF/LinkedData in future
The future
Project actively developed with the support
from JISC (UK) until January 2013
Basic maintenance covered until 2018
Google Ads to help cover the hosting costs
Participation on a future research projects
for improvements of the technology and
optimizations of the map management
Cooperation with libraries on improvement
Old Maps Online

www.oldmapsonline.org


      ... thank you ...

Edinburgh OldMapsOnline Workshop

  • 1.
    Who am I? Managingdirector of a small Swiss company: Klokan Technologies GmbH. Innovative geo applications in culture heritage. Software development services & customization of open-source projects, participation on grants. Clients: BL.uk, NLS.uk, David Rumsey (USA), NationaalArchief.nl, MZK.cz, Swiss libraries, ...
  • 2.
    Petr Pridal <petr.pridal@klokantech.com> www.oldmapsonline.org National Library of Scotland, Edinburgh, UK, December 13th 2012
  • 3.
    www.oldmapsonline.org The easy-to-use gatewayto historical maps in libraries around the world.
  • 4.
    Search engine formaps Library catalogs & fulltext search fail on maps. Search by geographic location / area on a modern world map Indexing maps from many different library collections One click to view the map in high resolution Free to access and open for contributions
  • 6.
    Contributor friendly Bring onlinevisitors to the websites of participating institutions to view the maps Promote the map libraries and their content Do proper crediting: logo and links back Store only minimal metadata & thumbnails Boost the number of online visits of maps Make it easier to discover the scanned maps
  • 7.
  • 8.
    More to come... Openfor new contributors. Institutions can let index their map collections. It is for free. The contributor retain all the rights. Ongoing negotiation with additional libraries. Submission of maps from individuals at a later point (technically more problematic)
  • 9.
    Contribute High resolution (300dpi+or 2 Megapixel+) maps with a zoomable viewer online Permanent URL with direct access - no password or payment required Metadata with DublinCore elements supplied in a spreadsheet (XLS). Later OAI- PMH harvesting. Bounding box (MARC 034, DCMI BOX, ...) http://project.oldmapsonline.org/contribute
  • 10.
  • 11.
    Bounding box encoding MARC: field 034 034: $$dE0055700$$eE0102900$$fN0474800$$gN0454900 255: $$cE 5°57'00"-E 10°29'00"/N 47°48'00"-N 45°49'00" DublinCore: DCMI Box DCMI BOX: westlimit=5.9559; southlimit=45.818; eastlimit=10.4921; northlimit=47.8084 ISO 19115 / FGDC / OGC
  • 12.
    How to createsuch records? BoundingBox Tool + Geoparser Tool Georeferencer Tool
  • 13.
    BoundingBox Tool BoundingBox Tool:Cataloguing while holding the physical maps - manual selection of the area on an online map (used by Swiss libraries, David Rumsey, ...) Geoparser: Semi-automatic rapid process - with suggestion of the area by analyzing existing textual metadata.
  • 14.
  • 15.
    Georeferencer Tool Enrichment ofthe online scanned maps A tool which runs in a web browser An online service - no software installation Uses volunteers for metadata enrichment Successful pilots in several institutions
  • 16.
  • 17.
  • 19.
  • 20.
    Georeferencer It allows toturn the existing online scans into geodata in a web browser Metadata enrichment, with 3D visualization as a side product Support true collaborative editing It retains log of all operations (a la Wikipedia) - you can adjust it with a few clicks Library can receive GeoTIFF / WMS / ...
  • 21.
    Project launched inthe end of February in New York and London Over 30.000 visitors per day on the website after launch Two servers in a load balancer to handle the traffic after launch. An amazing feedback from general public
  • 22.
    Publicity Social Networks: Twitter,Facebook Google Developer Showcase + Google I/O 2012 BBC History magazine online slideshow La Stampa (Milan, Italy) newspaper Gizmodo, Ars Technica, The Verge Wired.jp (Japan), About.com Genealogy, Playtech.r (Romania), Portsmouth News (UK), ... TV spot in Catalonya: link
  • 23.
    120.000 high-resolution historical maps covering the whole world (by the end of the project it is going to be close to 150.000 probably)
  • 24.
    During the last10 months we have helped over 400.000 people to find maps and visit online map libraries
  • 25.
    We have encouraged all libraries holding maps to think about their metadata in a geographical context.
  • 26.
    Technology Based on theMapRank Search technology http://www.mapranksearch.com/ Released in 2010 by Klokan Technologies for Swiss project Kartenportal.ch, improved for Czech libraries and for David Rumsey Useful for large metadata catalogs + geospatial digital repositories. Details in the article published in D-Lib: doi:10.1045/september2011-oehrli
  • 27.
    Technology MapRank ranking: anadvanced algorithm calculating coverage similarity of geographical area of query and area covered by every document and combining it with map scale, time range, predefined categories and traditional full text search.
  • 28.
    Technology Default user interfacedeveloped with Closure JavaScript Library + Closure Compiler. Server component: FastCGI, C++, runs on UNIX/ Linux Data indexing implemented specifically to provide ranked real-time results even for large databases. Traditional R-Tree indexing or Lucene are failing.
  • 29.
    with # ofrecords: 70.000 Performance 125.000 250.000 500.000 for 70.000 test queries 1.000.000 900 675 Time in seconds 450 225 0 old implementation new implementation
  • 30.
    Google indexing 1-3 positionon Google for “old maps” phrase The intuitive geographical search interface uses AJAX, it is only one page for Google bots :-( The project must expose the metadata in a way indexable by traditional full-text search engine Phrases such as the “Old maps of Edinburgh”, “Old maps of British Isles”, “of ...”, ... World coverage, high quality geodata required
  • 34.
    Linked network ofgeodata Natural Earth Data - possible export as RDF/LinkedData in future
  • 37.
    The future Project activelydeveloped with the support from JISC (UK) until January 2013 Basic maintenance covered until 2018 Google Ads to help cover the hosting costs Participation on a future research projects for improvements of the technology and optimizations of the map management Cooperation with libraries on improvement
  • 38.