250 likes | 412 Views
A Survey of Web Mapping Part 3: Linking DARMC and PLEIADES Guoping Huang, ghuang@cga.harvard.edu. http://darmc.harvard.edu. Present. Past. DARMC. Space/Location. Event/Feature. Reference. Time. http://pleiades.stoa.org. DARMC has: Name, location XY, source PLEIADES has:
E N D
A Survey of Web MappingPart 3:Linking DARMC and PLEIADESGuoping Huang, ghuang@cga.harvard.edu
Present Past DARMC Space/Location Event/Feature Reference Time
DARMC has: Name, location XY, source • PLEIADES has: PID(URL), Name, map sheet no., grid no., source (Barrington Atlas)
DARMC has: Name, location XY > sheet no., grid no. source • PLEIADES has: PID(URL), Name, sheet no., grid no., source (BA)
Challenges: • Challenges with place names: PLEIADES: typos in OCR, spelling variations DARMC: typos in digitization, original names replaced by commonly used alternative names
Challenges: • Challenges with place names: One name, many places
Challenges: • Challenges with sheet no.: Inserts
Challenges: • Challenges with grid no.: Geo-correction changed original location
Challenges: • Challenges with source: Some original BA points replaced by TIB points
Grid no. Name Sheet no. Source PLEIADES ID
Grid no. Street no. Name Street name Sheet no. Zip code Source City State PLEIADES ID X,Y
Advantages : • Alternative place names (Alias table) • Fuzzy match: allow spelling variations (Spelling sensitivity) • Interactive match (Candidate score)
Results • 24,030 Roman places/points in DARMC 6,171 points don't have names 20,995 points are indicated as from Barrington Atlas • 14,922 points are matched with PLEIADES ID = 84% of named Roman places in DARMC