Statistical Analysis

1
Statistical Analysis Dissemination of Census
Data
2
Statistical Analysis and Dissemination of Census
Data

• Outline
• The Power of Maps
• Introduction and Example
• Dynamic Census Atlases
• Overview Example
• Spatial Analysis Techniques
• Overview Examples
• Digital Geographic Data for Dissemination
• Overview Cost and Benefits
• Digital Data Dissemination Strategies and Users
• Overview of Users

3
Anyone or anything can be associated with a known
location in the world
4
CHILE HOUSING AND POPULATION CENSUS DISTRICTS
2002
5
Tsunami Affected Areas in Gizo, Solomon Islands
6
The power of maps

• Maps communicate a concept or an idea.
• Maps are often meant to support textual
  information
• Maps appeal to the viewers curiosity
• Maps summarize large amounts of information
  concisely
• Maps can be used for description, exploration,
  confirmation, tabulation
• Maps encourage comparisons
• Between different areas on the same map where
  are population densities highest?
• Between different maps is child mortality higher
  in the districts of province A than in province
  B?
• same area where and by how much do literacy
  rates for males and females differ in the
  districts?
• Between maps for different time periods did
  fertility rates decline since the last census?

7
Dynamic census atlases

• Alternative to a static census atlas
• Publishing of a digital map and database together
  with mapping software can allow users to produce
  custom maps of census indicators.
• Normally includes digital boundary files at a
  lower resolution than the full census database to
  allow fast drawing and low disk usage
• closely integrated attribute table should contain
  only a selected number of census indicators.
• Densities and ratios that are appropriate for
  mapping should already be calculated.

8
Dynamic census atlases

• The data provider should therefore provide an
  easy-to-use package together with the boundaries
  and data.
• The use of that package should require minimal
  training and experience.
• The application should be plug-and-playafter
  installation, the user should immediately be able
  to produce maps
• Drill-down options for different geographic
  selections
• Interactive area delineation options (e.g. select
  schools in a district)

9
A screenshot of Ukraines dynamics census atlas
10
Spatial Analysis Techniques

• the main use of spatial analysis is for census
  products and services
• Techniques include buffering, linear
  interpolation, point pattern analysis, and
  cartograms, etc.
• All offer functionality beyond standard thematic
  (choropleth) mapping, with many tools now
  available in both commercial and open-source
  software programs.

11
Spatial Analysis Techniques

• Some prevalent forms of spatial analysis
  especially useful for use with population data
  include
• Queries
• Distance measurements
• Transformations
• Buffering
• point-in-polygon analysis
• Polygon overlay analysis

12
Spatial Analysis Techniques

• Queries
• Often this is the first step in an analysis,
  where one seeks to create a subset of units such
  as populated places with certain characteristics,
  allowing the user to check how typical an
  observation is against other observations
• They use a GIS program to answer simple questions
  posed by the user, with no changes in the
  database and no new data produced.
• An example of a query using geocoded census data
  is, select all towns with a population greater
  than 1,000 persons. These towns can then have
  their attributes summarized, for instance, to
  measure their total fertility rates against
  smaller towns and villages, then map the results
• The term exploratory data analysis refers to
  investigations of patterns and trends in data
  using such techniques as querying

13
Area delineation

• E.g. Interactive determination of school
  districts with the same number of children in
  each school grade by aggregating census
  dissemination areas

14
Spatial Analysis Techniques

• Distance measurements
• Easily done with all GIS programs, using the
  centroids (or center points) of cities, towns,
  and villages.
• An analysis can be done to select villages
  located more than a kilometer from a school,
  clinic, or water source.
• These can then be further analyzed using the
  attribute information for the populated places
  themselves.

15
Spatial Analysis Techniques

• Transformations
• Methods of spatial analysis that use simple
  geometric, arithmetic or logical rules to create
  new datasets
• Transformations can include operations that
  convert raster into vector data, or a stream of
  GPS coordinates into a route or a boundary
• Of all the transformational techniques, buffering
  is the most well known and important

16
Spatial Analysis Techniques

• Buffering (transformation)
• Involves building a new data layer by identifying
  all areas that are within a certain specified
  distance of the original.
• Buffering can be performed on points, lines and
  polygons and can be weighted by attribute values.
  
• Buffering can be used to model travel time, for
  instance, by creating a catchment area around a
  particular feature such as a school or a clinic.
• This provides a measure of accessibility that can
  be mapped across the extent of a country.

17
is near to Buffer Operations

• Point buffer
• Affected area around a Hospital
• Catchment area of a water source

18
Buffer Operations

• Line buffer
• How many people live near the polluted river?
• What is the area impacted by highway noise?

19
Buffer Operations

• Polygon buffer
• Area around a reservoir where development should
  not be permitted

20
Spatial Analysis Techniques

• point-in-polygon analysis
• Determines whether a point lies inside or outside
  a polygon.
• Can be used to compare geocoded village centroids
  lying inside and outside hazardous areas such as
  tropical storm tracks or earthquake zones.
• Polygon overlay analysis
• Involves comparison between the locations of two
  different polygonal data layers.
• For example, the boundaries of two administrative
  districts could be compared to troubleshoot
  errors in the field enumeration process

21
Spatial Analysis Techniques

• Spatial interpolation
• A spatial analysis method designed to fill in
  values that lie between observations
• A variety of methods including inverse-distance
  weighting and kriging are used to estimate the
  values of unsampled sites
• based on Toblers first law that all nearby
  objects are more similar than distant objects
• Kriging interpolation technique for obtaining
  statistically unbiased estimates of spatial
  variation of known points such as surface
  elevations or yield measurements utilizing a set
  of control points
• In kriging, the general properties of a surface
  are modeled to estimate the missing parts of the
  surface

22
Example of linear interpolation creating contours
23
Thiessen polygons illustrated
Spatial Analysis Techniques

• Thiessen polygons
• Have the unique property that each polygon
  contains only one input point (e.g. settlements),
  and any location within a polygon is closer to
  its associated point than to the point of any
  other polygon.
• This method assumes that the values of the
  unsampled data are equivalent to those of the
  sampled points.

24
Areas of influence

• Commuting distances daily commuters flow

25
Spatial Analysis Techniques

• Descriptive summaries are a spatial equivalent of
  descriptive statistics (such as mean and standard
  deviation) that represent the essence of a
  dataset in 1 or 2 numbers
• Centers of population are the two-dimensional
  equivalent of a statistical mean and are often
  used to display the center of population using
  the weighted average of x and y coordinates of
  populated points
• Point pattern or cluster analysis regards the
  distribution of points in space irrespective of
  their actual locations to determine whether
  patterns are random, clustered, or dispersed
• hot spots are where high values are surrounded by
  high values, or cold spots, where low values are
  surrounded by low values. These are particularly
  useful for identifying populations at risk as
  well.

26
Spatial Analysis Techniques

• Cartograms
• sometimes used to display census results
• The areas of the original polygons are expanded
  or contracted based on their attribute values
  such as population size or voting habits

27
Modelling smoothing

• Evolution of the population beetwen two censuses

28
Digital Geographic Data for Dissemination

• Demand for digital databases that consist of
  extractions of the census agencys digital
  geographic master database will only increase
• Census data are an important input in policy
  planning and academic analysis in many fields.
• Health service provision, educational resource
  allocation, design of utilities and
  infrastructure, and electoral planning are some
  applications where government agencies require
  spatially referenced small area population
  statistics.
• Commercial users employ such data for marketing
  applications and location decisions.

29
Digital Geographic Data for Dissemination

• Benefits and costs
• Benefits Unsurpassed detail and precision, the
  potential use of census data in numerous
  applications--especially when overlaid on other
  geographic data such as terrain, and the relative
  ease of management and storage of thousands of
  units
• Costs expense in processing and data management,
  possible data disclosure issues, and quality
  controlcosts of metadata production should be
  factored into the equation as well

30
Digital Data Dissemination Strategies and Users

• The wide range of potential users of
  disaggregated census data means that the NSO
  needs to pursue a multi-leveled digital data
  dissemination strategy.
• Broadly, we can distinguish between the following
  types of users
• Advanced GIS users
• Computer literate users
• Novice users

31
Digital Data Dissemination Strategies and Users

• Advanced GIS users
• work easily with large datasets and can use ftp
  to access them
• Require extensive metadata. Sometimes called data
  extractors or power users
• They will want access to spatial and attribute
  information in a comprehensive digital geographic
  format
• The census office needs to supply comprehensive
  documentation on the geographic parameters used
  for the geographic database as well as on the
  individual census variables
• The spatial information will be distributed in an
  open geographic format that can be easily
  converted into any number of commercial GIS
  formats

32
Digital Data Dissemination Strategies and Users

• Computer literate users
• Government, commercial or private sector users
  who want to be able to browse the thematic
  information in a census database spatially.
• Want to produce thematic maps and thus need to be
  able to perform simple manipulation of
  cartographic parameters.
• Simple analytical functions such as aggregation
  of census units to custom-designed regions should
  also be possible.
• This group of users is best served with a
  comprehensive, pre-packaged application that is
  designed for a commercial or freely available
  desktop mapping package.
• Documentation requirements are somewhat smaller,
  since the users are unlikely to change the
  geographic parameters of the database or perform
  more advanced GIS operations.

33
Digital Data Dissemination Strategies and Users

• Novice users
• Largely want to view pre-prepared maps on a
  computer and perhaps perform some basic queries
• Best data distribution strategy is often to
  produce a self-contained digital census atlas
• This atlas could consists of a series of static
  map images, for example, in the form of a slide
  show
• Or it could be a very simple mapping interface
  with pre-designed map views that allow basic
  queries
• Both, static maps and a simple map interface, can
  be made accessible through the Internet

34
GRACIAS POR SU ATENCIÓN