Lecture two

1
Lecture two

• Storing geographical data

2
This lecture

• How do we store geographical features?
• Types of data.
• Vector data storage.
• Layers, Coverages and Workspaces.
• Using ArcCatalog.
• GeoDatabases.

3
The grim pasthow we used to do it

• Geographical markers
• Rome boundary stones were sacred to Jupiter
  Terminus, to whom you were given as a blood
  sacrifice if you moved one.
• Yorkshire cup and ring marked rocks.
• Communal reemphasis
• Rogationtide in English villages beating the
  bounds.

4
Maps

• Maps been around since at least 6200 B.C (Catal
  Hyük town plan in Turkey).
• Forma Urbis Romae 3rdC. Marble map of Rome
  including every room. 17x12m.
• Usually made for taxation.

5
Now

• The public, cheap availability of maps has
  removed the need for communal reemphasis of
  boundaries, except in Northern Ireland.
• Still use geographical markers, but not for
  anything serious other than navigation.
• No one gets killed for shifting a boundary!
• All boundaries used for any governmental,
  academic or business use are kept in a GIS.
• Of these, the greatest number are in ArcGIS
  format.
• So, how does ArcGIS store maps?

6
ArcGIS Data Models

• Old ArcInfo model Coverages and Feature files.
• Works in both ArcWorkstation and ArcDesktop.
• New ArcGIS model GeoDatabases.
• Works in ArcDesktop.
• Can also use
• Shapefiles ArcView files traditionally. Useful
  for moving data about.
• Image files Raster, again, useful for moving
  data.

7
The General Data Model

• We break data up into its constituent parts or
  Features.
• We can do this splitting with most
  human-influenced data (eg. Towns, forests,
  boundaries etc.).
• Commonly ArcGIS deals with line and point data
  Vector data.

8
ArcGIS Basic Vector Types

• Label points (single x,y coordinate locations)
• eg. phonebox on an Ordinance Survey (OS) map
• eg. a town on a global map.
• Arcs (lines of multiple x,y coordinates)
• eg. roads, rivers.
• Polygons (enclosed areas made from Arcs)
• eg. house plans, parks, boundaries.
• Attributes (associated non-location information)
• eg. census data for and area, house names,
  population numbers.

9
Other data

• Some data doesnt easily break up into lines and
  points
• eg. satellite images, terrain surfaces, photos.
• This is often handled as Raster data. At its
  simplest, a raster file contains one value for
  each pixel in an image.
• Obviously not very efficient to store compared
  with end points for a vector arc.

0,10 100,90
10
Other data

• Some terrain and other surfaces can be
  represented by lines.
• Done using a Triangulated Irregular Network or
  TIN.
• You can also use Lattices - regular grids of
  points, each with a height that makes an overall
  surface.
• Well look at non-Vector data in the coming
  weeks. For now well look at how Vector data is
  stored.

11
The ArcInfo Data Model

• Store all the same types of features in one file
  using x,y coordinates
• i.e. all the point features in one file, all the
  line features in another file, and all the
  associated non-location data in yet another file.
  
• Makes storage and handling easier.
• Makes analysis easier.
• The filenames show what kind of data they hold.
• e.g. Label Points are stored in a LAB (lab.adf)
  file.

12
Storing Vector data

• The files hold coordinate data and Feature
  Numbers that are assigned sequentially to
  Features in the file.

13
Tics

• Tics geographical control points used to
  register (overlap) different datasets, and
  transform (eg. put in a new projection scheme)
  the data.
• For example, you might mark in known locations on
  two datasets, usually at the data edges, as Tics.
• Stored in a TIC file (tic.adf).

14
Arcs

• Arcs - stored in an ARC file (arc.adf).
• A point where an Arc changes direction is called
  a Vertex.
• The start and ends of Arcs, and the crossing
  point between two Arcs, are called Nodes.
• The ARC file stores each Vertex and Node as an
  x,y coordinate.

Data can be assigned to Nodes
15
Polygons

• An area - can also have islands inside, that
  cut out inner areas.
• Made from one (or more) Arcs.
• The end Node is the same as the starting one.
• Each also has a Label Point (x,y) associated with
  it, stored in a LAB file.

42
16
Polygon-Arc List

• The Polygon-Arc List file (pal.adf) stores which
  Arcs make up each Polygon. That way we dont need
  to store the Arc coordinates again.

ARC
PAL
1
2
23
3
17
Polygon-Arc List

• The outside or Universe Polygon is always
  feature number 1, ie. first in the PAL file.
• Islands start with a zero, followed by the Arcs.

18
Topology

• Analysis we might want to do
• How do I move from Arc to Arc to get from A to B
  fastest?
• If I leave Polygon A going north, which Polygon
  do I enter?
• To do these analyses we need some notion of
  Topology, ie. the spatial relationships between
  features.

19
Topology in ArcInfo

• Three ways topology is defined in ArcInfo
• Arcs connect at Nodes.
• Arcs have direction, and therefore a left and
  right side.
• Arcs that connect to surround an area make a
  Polygon.

20
How Topology is stored
Note that polygons are listed clockwise in the
PAL, with minus figures for reverse direction
Arcs.

• Stored in the files.

N2
PAL
23
N1
N3
ARC/ATT
21
Feature Attributes

• Info stores information about each feature in a
  Feature Attribute Table using the Feature Number
  to match the attribute data to the coordinates in
  the other files in a one-to-one relationship.
  Each data type has its own table type.
• Each feature will also have a unique User Defined
  ID kept in its table in addition to its Feature
  Number.

PAL
FAT
22
Types of Feature Attribute Table

• Label Points / Polygons Point/Polygon Attribute
  Table (PAT / pat.adf file).
• Includes AREA and PERIMETER columns with the
  values associated with any Polygons for which the
  Points are labels.
• Can be used to hold points or Polygon labels, but
  not both.
• Arcs Arc Attribute Table (ATT / att.adf file).
• Includes FNODE, TNODE, LPOLY, RPOLY, LENGTH
  columns.
• When Nodes have data associated with them - Node
  Attribute Table (NAT / nat.adf file).
• Includes an ARC column referencing one of the
  Arcs.

23
Feature Attribute Table Format

• After the columns listed above usually come the
  Feature Number (NAME) and the User Defined ID
  (NAME-ID) columns.
• Following these, users can then create more
  columns to hold attribute data.

24
Example PAT

• Note that the Universe Polygon is first, and the
  AREA given for it is the negative total of the
  rest.

25
Other Vector data

• Region several related Polygons, nested or
  overlapping. Stored in a PAT.regionName (pat.adf
  file).
• Annotation text drawn along an associated
  feature. Stored in a T(ext)AT (txt.adf file).
• Section an Arc / portion of Arc representing a
  part of a pathway. Stored in a SEC table (sec.adf
  file).
• Route a pathway made out of multiple Sections.
  Stored in a R(oute)AT (rat.adf file).

26
Putting features together

• So, weve seen that individual features are
  stored in files with similar types (e.g. all the
  line features in a file of Arcs).
• When these are combined, you get a data
  Coverage. A Coverage usually contains one type
  of geographical information or analysis result,
  e.g. Geology, Roads, or Quickest routes
• Several Coverages may go together in a map, and
  you can turn different ones on and off to display
  different facets of a map.

27
Coverages

• In the file system, these are represented as the
  directories your data files go in.
• They let you keep data files together and display
  them at the same time.
• Usually it is the name of the Coverage that goes
  to form the Feature Number and ID column names.
• E.g. The SOILS coverage before gave our example
  PAT a SOILS Feature Number column and a SOILS-ID
  column.

28
Addition Coverage information

• As well as our data files, a Coverage will have
  several other pieces of information associated
  with it.
• A Coverage Extent (BND / bnd.adf file) - this
  records the upper right and lower left corners of
  a rectangle containing all the feature data in
  the Coverage. It need not contain all the Tic and
  Annotation points.
• A Coordinate Definition file (PRJ / prj.adf ) -
  holds the Coverages map projection information
• A Tolerances file (TOL / tol.adf ) - holds a
  number of processing tolerances, eg. How close
  Tics in different files need to be before they
  count as matching.

29
Putting Coverages together

• Finally Coverages can be kept together in project
  areas called Workspaces.
• Workspaces allow you to keep all the data and
  Coverages you generate for a project in one
  place, separate from any other work you may be
  doing.
• In the file system theyre represented as
  directories which include all the Coverage
  directories youre working on in a project.
• One of the first things you do on starting any
  ArcWorkstation project is move to your Workspace
  using the w command.
• w m\workspace\schools

30
The Vector Data Model

• Different views of the same data.

File system
ArcCatalog
Workspace
Coverages
Vector data
Raster data
TIN data
31
Summary

• Vector data includes Label Points, Arcs, Polygons
  and Tics.
• The coordinates for the features of each type are
  stored in separate files, one file for each type.
• Within the files, each individual feature has a
  Feature Number and a set of coordinates.
• The Feature Number is used to link features to
  attribute data stored in a Feature Attribute
  Table of a particular type.
• You can get at the combined data through their
  representation as Tables in ArcWorkstation.

32
Summary

• Data stored in separate files / tables can be
  combined to form Coverages.
• Coverages usually contain information about a
  single landuse or analysis result.
• All the Coverages in a project can be kept
  together in a Workspace.
• Coverages and Workspaces are represented as
  directories in the file system.

33
ArcCatalog

• It is vital that you never alter the file names
  or location outside of the ArcGIS applications.
• You can get at most of the information in the
  files and change/move/delete them in ArcGIS.
• The best way to do these operations is in
  ArcCatalog.
• We looked at some of the functions of ArcCatalog
  in the last practical. Well add a bit more
  detail now.

34
Navigating

• ArcCatalog starts up displaying your harddrive,
  but you can add additional drives.
• You can add online databases if they hold
  appropriate data.
• Databases are a whole other course! Check
  ArcCatalog Help files for more details!

35
Data viewing

• Can view Coverages in a number of ways.

36
Metadata

• Data about Data.
• Where its from and whats happened to it.
• Who made it and whats its problems.
• What does it contain, and how.
• ArcCatalog lets you see and change metadata for
• Describing the data and where its from.
• Describing the spatial data form and projection
  etc.
• Describing the attributes associated with the
  data.
• Click on a metadata element to open up the
  metadata form. Edit, import etc. buttons above
  viewer.
• The metadata is in the Coverages metadata.xml
  file.
• XML is a markup language like HTML. You can open
  it in Internet Explorer if you want to see what
  it looks like.

37
Altering and copying Coverages

• You can use ArcCatalog as if it were Windows
  Explorer to rename, delete, copy or move
  Coverages and other data items, and make new
  folders.
• You use exactly the same keys/menus as you would
  in Explorer. Most menu options can be got at by
  right-clicking on a folder or file.
• The only difference is that ArcCatalog
  simultaneously goes through and updates all the
  files ArcGIS needs to keep track of whats
  happened to the data.

38
Making a Workspace

• Making a new Workspace is as easy as making a new
  folder.
• File gt New gt ArcInfo Workspace
• (or right-click gt New gt ArcInfo Workspace)
• Rename the Workspace appropriately.
• Note This fails if you have a space in the path
• If you look in Explorer, youll see an Info
  directory has been made in the Workspace
  directory to store related information.

39
Making a Coverage

• Making a Coverage is almost as easy, but requires
  slightly more thought.
• File gt New gt Coverage
• This brings up a Wizard (set of instructional
  forms to fill in) to help you.

40
Name the Coverage

• You can use an existing coverage to supply
  boundary, tic and projection information.

41
Fix the projection

• If you dont supply a template, youll need to
  say what projection scheme its in, or pick none.

42
Generate initial topology

• If you know which feature type will be important,
  you can generate the appropriate feature table.

43
Other ways of manipulating files and Workspaces

• Command line from Arc.
• Using any of the
• ArcTools.

• ArcTools is a set of AML routines with menus for
  doing tricky Arc tasks.

44
Summary

• Never alter the files or directories outside of
  ArcGIS.
• The easiest way to alter, delete or move ArcGIS
  data is with ArcCatalog. It acts like Explorer,
  but informs Arc what its doing.
• You can manipulate data and directories using Arc
  or ArcTools.

45
Geodatabases

• New ArcDesktop data format.
• The future for ArcGIS.
• Like a Coverage only stored in an Object
  Orientated Database.
• Local or Personal Geodatabases vs. SDE
  Geodatabases shared on a network.
• SDE databases can be on ArcSDE servers or most
  other databases. They have locking and versioning
  to resolve conflicts if more than one person
  edits a Feature.

46
Geodatabases

• Can build them from scratch, import from other
  formats or build using drawings.
• Keep strict Topological and Data Item constraints
  on Features.
• Build Topology and confirm Features have
  appropriate data as they are edited.
• Because of this, are used as the basis of Network
  Analysis.

Geodatabase
Feature Dataset
Feature Class
Features
Workspace
Coverage
Arcs, Polygons etc.
Features
47
Object Orientated Databases

• Each feature is an Object stored in the
  Geodatabase.
• Each Feature is of a certain predefined type, or
  Class.
• You can define your own classes, and inherit
  properties from standard classes.
• When your classes inherit from another class,
  they pick up all the Classes Attributes.
• Classes inheriting from another are known as
  Subclasses of a Superclass.

ESRI Simple Edge Feature integer length
Building integer people
House
Has length and people
Also has length
48
Constraints

• Geographical.
• Class and data types.
• Variable ranges or categories.

49
Example

• Whereas previously youd make an Arc for a road,
  you can now make a Road class from the Simple
  Edge Feature superclass.
• Constrain the Road so it must have a Surface type
  variable picked from a list.
• Subclass Road to make a Motorway class, with a
  Number variable which is constrained between 1
  and 999.

50
Building a Geodatabase from scratch

• First decide your requirements. Many parts of the
  creation process cannot be undone later, so get
  it right first time.
• What types of data do you want and what
  constraints?
• What geographical area do you need?
• Are you going to do analysis on the data?
• Build the Geodatabase in ArcCatalog.

51
Building the Geodatabase

• Right-click on the Workspace and pick New
• Unless a server is running ArcSDE, youll only
  get the chance to create a Personal Geodatabase.

To connect to an existing SDE database, use
Database Connections.
52
Building the Geodatabase

• You set constraints by creating Attribute Domains
  at the Geodatabase level in its Properties.
• These can be ranges or coded values (categories)
  data must fall in.
• These are then applied to specific Feature
  classes.

53
Attribute Domains

• Maximum and Minimum range, or build codes for
  categories with associated text descriptions.
• A Default Value for new Features.
• Split policy if a Feature is split, how should
  the Attribute data be handled in each new
  Feature?
• Each gets the default, duplicates the original,
  or is the original value split on the basis of
  the divided length/area.
• Merge policy if two Features joined, how should
  the Attribute data be kept?
• Feature gets the sum of the original, the
  default, or a length/area weighted average.

54
Feature Dataset

• Right-click Geodatabase gt New gt Feature Dataset.
• A collection of Features that share a Spatial
  Reference, i.e. area and coordinate system. Set
  this with the Edit button.

Give the Dataset a max/min area and Select a
Projection System.
55
Feature Classes

• Right-click Feature Dataset gt New gt Feature Class
• Choose a name and set the field.

56
Feature Classes Constraints

• Add new Field.
• For each pick an Alias (alternative name - can
  include spaces), a Default Value, and an
  Attribute Domain if you want constraint.
• Note that Long Integer Integer used when making
  domains.
• Field Properties change when you click on a Field
  Name.

57
Table of Contents

• You should end up with something like the
  following.

Geodatabase
Feature Dataset
Feature Class
58
Drawing your Geodatabase

• You can also make your Geodatabase schema (i.e.
  structure) using a combination of
• CASE (Computer Aided Software Engineering) tools.
• Allow you to draw software and convert it to
  computer code.
• UML (Unified Modelling Language).
• A standard way of drawing Object Orientated
  things.

ESRI Simple Edge Feature integer length
Occupants text owner integer others
Building inherits from ESRI Simple Edge Feature
Building integer people
Building has attached Occupants data
59
Using UML

• ArcDesktop comes with a copy of Visio - CASE
  software for drawing UML.
• Draw your classes, defaults etc. as UML diagrams.
• Export from Visio into a Microsoft Repository
  (e.g. Access) to convert it into Microsofts
  Object Orientated database format.
• Import the schema into ArcDesktop.

Youll need to add in the CASE tools.
60
Summary

• Geodatabases are a new, highly controlled data
  storage technique.
• When you build one, you can set Attribute Domains
  which are constraints that can be applied to
  Features.
• You build a Feature Dataset for a particular area
  / spatial reference.
• You design Feature Classes e.g. Roads and set
  their constraints.

61
Next Lecture

• Getting data into ArcGIS.
• Getting data into the ArcInfo data model.
• Getting data into GeoDatabases.
• Creating topologies.
• Getting data out of ArcGIS.