Geog 495 Final Exam Review 112305 Final exam

Geog 495 Final Exam Review 11/23/05 Final exam on 11/28/05 Mon

Outlines I. Big picture – Attribute – Space – Attribute & Space II. Small picture – GIS view on spatial data – DBMS view on spatial data III. Synthesis

I. Big picture • Geographic information has three main components: space, time, and attribute • Traditionally the representation of attribute is well studied in contrast to space and time • How attribute is represented in computer database model (first half of this course) • How space is represented in computer spatial data model (second half of this course) • How space is combined with attribute – In GIS (GIS Architecture) – In DBMS (SDBMS Architecture) – Integration of GIS with DBMS

Database model • • • Hierarchical Network Relational Object-oriented Object-relational

Spatial data model • Object view • Field view

GIS architecture • Hybrid • Purely relational • Object-oriented or object-relational File-system into database system

SDBMS architecture • Purely relational • Object-oriented • Object-relational Accommodated into spatial semantics

Integrating GIS with SDBMS • GIS-centered – e. g. Arc/Info ODBC – What’s the relative advantage of this approach? • SDBMS-centered – e. g. Oracle Spatial – What’s the relative advantage of this approach?

GIService • Utilize specialized GIS functionalities that meet specific needs • Built upon DBMS • Somewhat eclectic: good things from both

II. Small picture • Two views on spatial data model • GIS view: focused on how to store spatial information suited to application needs (e. g. building topology for routing applications) • SDBMS view: focused on how to store spatial information suited to SQL-like queries (e. g. defining spatial data type, spatial operators)

1. GIS view on spatial data • Data hierarchy Human thinking – Data model: how spatial concepts are viewed – Data structure: how spatial information in stored in computer Level of abstraction – Data format: how data structure is stored in s/w specific format Machine code

Data model • There are two common views • Object-view – The world is composed of discrete entities • Field-view – The world is composed of continuous fields • Both coexists in the real-world • One phenomenon can be represented in two views • Human perception tends to discretize information also

Data structure • Data structure that stores spatial information – Vector: composed of (a set of) point – Raster: composed of grid cell • Data structure that stores attributes derived from spatial relationships – TIN: connected network of points whose attributes vary – Matrix: represent attributes based on a pair of spatial objects

Do you know? • Vertice, node, point, line, polygon, multishape polygon, multistring…. • Topology vs. spaghetti model • Planar vs. non-planar • Compression methods for raster data structure – Run-length code – Quadtree

File format • Different systems use different file formats • Needs for interoperability – How can we promote interoperability? • Metadata – Metadata standard? – What kind of information is documented? – FGDC standard

Do you know? • • SDTS XML Open GIS Characteristics of specific file format and how they are classified into different data model? – TIGER/Line – DEM – Mr. SID, BIL, Geo. Tiff (These are common DOQ format)

2. DBMS view on spatial data • Spatial data type • Spatial operators

Spatial data type • Object – Point sets – Do you know OGIS spatial object types? • Field – Tessellation (grid) representation – No too much work has been done

Spatial operators • On field-based data – Local (e. g. +, -, …) – Focal (e. g. gradient) – Zonal (e. g. average) • On object-based data – Topological (e. g. within, overlap, touch) – Metric (e. g. distance, direction, area)

Miscellaneous • What is spatial access method (or spatial indexing)? – Quadtree vs. R tree • Query processing – Single scan vs. multi scan

III. Synthesis • More DBMS approach to GIS is desirable – DB system is superior to file system – DB gets increasingly smarter – Distributed GIS environment gets increasingly popular

• Inadequacy of RDB to representing spatial concepts should be accommodated – User-defined spatial data types – User-defined spatial operators – Custom data model (refer to next week presentation on geodatabase case studies) – Geographer’s role formalizing spatial concepts (Naïve geography? )