Total Pageviews

Sunday, 24 November 2013

Implementing GIS to tackle local government issue - Part 5

Part 4: http://daneshatlas.blogspot.com.au/2013/11/implementing-gis-to-tackle-local_8447.html


Conceptual Design

Data design

The key data element of address dataset that would be logically linked with community facility table would be EZI_ADD. The reason this element is crucial is because EZI_ADD is used to geocode the facility address. For this linkage to be successful, the addresses of the community table must be identical of format and address representation.

The minimum map area is 0.01 hectares, based on CCD .The error tolerance for this project is 10%, to match the error tolerance of dataset’s positional accuracy. The map scale desired for the project would be 1: 1000. The map projection that should be employed for the project area is MGA Zone 55 as this projection minimizes distortions for the region. Temporal resolution would be every year to cater annual needs of the city.  

Data sources for this information product come from DSE (public source). The lineage of address dataset starts from derivation from address information found in Vicmap Property and Vicmap transport. The addresses were later analysed and categorised and channelled the results to address dataset. This address dataset has address match rate of 92% (2008/09). The data standard I will employ in this project would be ANZLIC Metadata Product Guidelines VER 1.2 July 2011. The reason I chose this standard is because the metadata format for the DSE datasets are based on this and applying the same standard would fasten the process of writing product description. Since the datasets found are readily available in ESRI shapefiles, there will no need for data format conversion.

Logical Data Model

Most or all the dataset structure for the project would be vector-based. Below-here is an example the characteristics of vector transportation dataset.
Fields
Data Type
FID
Object ID
Shape
Geometry
PHYS_COND
Text
ROTATION
Float
ROAD_REL
Text
FTYPE_CODE
Text
CRDATE_PFI
Date
PFI
Long Integer
NAME
Text

The suitable structure for the data is object-relational data model. It allows addition of specialized behaviour to the relational data model which is close to real world behaviour.
The advantages of this model would be it maintain high data integrity, multiple users are able to update simultaneously and allows use of legacy and GIS databases. For the project, this model is chosen because ArcGIS software uses this model and it allows wide range of relational and objects rules to be applied. Disadvantages would be it is compromise of two previous models and difficulty of modelling complex relationships.

For attributes to be linked with the administrative dataset, the geo-relational data model is most appropriate model for storage and manipulation of the data. This is because the parcel boundaries are able to be linked easily with other tables containing data of population characteristics.

The model (incorporating relational and objected orientated models) allows relational data model to have object-orientated extensions. This means various relational and object-orientated rules are applicable which are needed for the project dataset (parcel layers and network datasets). Secondly, its ability to allow multiple, simultaneous updates would maximise GIS analyst working efficiency. For example, one GIS analyst would be using series of datasets to analyse accessibility to clinics while another analysts would use the same datasets for another analysis.


Part 6 (Final): http://daneshatlas.blogspot.com/2013/11/implementing-gis-to-tackle-local_8847.html

No comments:

Post a Comment