• Randmark 40
  • Randmark 40
 
Kevod Insurance System
 
KS4 Data Platform

 

RM40's postal dataset Print E-mail



 

Everything happens Somewhere!

An address might not always be specified, but risk is always limited to a specific address, a country, a region, or a planet.  Neil Armstrong’s domestic policy would not have covered loss of personal effects during his walkabout on the moon! Addresses form an important foundation of insurance. Much is dependent thereon whether it is a risk item, an insured event, or the rating factor of a premium.  

In South Africa, a postal code is a four-digit number that assists the Post Office with the automated sorting of mail. Each code is linked to a post office, a postal agency, a postal depot or a mail collection point. 

In SA, there are more than 58 000 place names, 10 000 postal codes,  and an approximate 16 000 place names have postal codes ascribed to them. Moreover, place names change : "Pietersburg" has become "Polokwane", "Messina" became "Musina", "Ellisras" became ""Lephalale". Certain place names have bilingual spelling, for instance: “Die Wilgers” is often referred to as “The Willows”. "Cookhouse" is also known as "Kookhuis". In addition, the same code can be used for multiple place names, and the same place name can have multiple codes associated with it! One must remember that the postal code list subtly changes from time to time - apart from "growing" ever-so-slightly every month - as  postal codes were not intended as absolute spatial references.

Allthough a postal code links a placename, sorting centre and province to one another, the provincial demarcation does not  correspond with actual boundaries, but postal hub zones.  It is not always clear what the "town name" linked to the code constitute, as one finds "Johannesburg" or"Pretoria" on thousands records, however, neither of them being a town. 

In addition to the above, the public often do not know what the relevant postal code of an address is. We may mistake our postbox code as our area code, or might assume that we live in Hyde Park whereas we actually live in (next-door) Dunkeld West. Garbage in, garbage out...

However - and despite the above - postal codes are used as a spatial references  that are usefull, widely used and quite effective. Even CRESTA (Catastrophe Risk Evaluating and Standardizing Target Accumulations) ZONE codes that describe natural-disaster risk areas, are assigned to postal code demarcation. 


What have we done? 

A very useful dataset is contained in the South African Magisterial District demarcation, as it forms the basis of: 

a)   general census data,

b)   crime statistics, 

c)   petrol price zoning.

These districts constitute stable data points, and a wealth of data is available for comparative purposes. 


We have tackled the existing list/s head-on, and compiled our own list of place names, with: 

a)   place name duplications removed i.r.o. Eng / Afr spelling, 

b)   alternative (old) place names kept intact (for searchability), 

c)   new place names and/or spelling format included , 

d)  new postal codes updated, 

e)   CRESTA zone codes (primary & secondary) incorporated, 

f)   CRESTA codes aligned with post - 1994 provinciual demarcations,

g)   Post Office hub zones included,  

h)   postal codes having been resolved to magisterial districts,

i)   the Post Office's list's 'town name' references reconstituted as 'area name', having improved on it,

j)   referencing " known as" references to placenames (in process). 

 It simply means that when you select an address, our Ks4 system populates a substantial number of additional fields associated with your selection, which becomes available for report purposes. Potentially you could track claims loss ratio's per magisterial district, and interpolate such data with census statistics ; as an underwriting tool the list is an absolute breath of fresh air! In due course it might become the foundation of all insurance reports. It's - quite probably - simply as "good as it gets".

We believe the RM40 postal code dataset to be the most cost effective spatial dataset tool available, today. It is incorporated into our Ks4 insurance data platform, but is also available as a standalone, software-as-a-service arrangement, whereby the list is made available, and updated (quarterly). 


What do users of the RM40 postal dataset have to say about it?

  •  “the RM40 postal dataset is a most valuable resource in linking geographic information to ...postal codes “.
  • “a huge amount of value and benefit to (any) provider would be...  in the entry of information at source ... so that information is entered right (and matched to postal code) first time.”
  •  "User" data is rifled with challenges such as non-matching of postal codes to placenames, randomly contracted place names (KTown, K's Town, KWilliams Town, KWT, King William's Town, King Williams Town), non-combined options (Port Elizabeth / Nelson Mandela Bay / P.E., NM Bay, or PE / NM Bay) with unclear user matches, typos, missing data (place names / postal codes) : We now have the magisterial districts databased and coded, and we can do an address-by-address lookup to find the municipal district via a simple query.
  • “the cross-reference to postal code is invaluable to validate the address, and gives greater certainty to results.”
  • “as a source of clean data inputs ... the RM40 postal dataset is ... very useful in data clean-up and validation, particularly in the cross-mapping of postal and residential codes.  It's even a good initial / supportive stepping stone to a move to geo-coding addresses”.
  • "We have linked place name GIS coordinated from a data base at Berkley, and have constructed a spatial envelope algorithm to measure distance of an individual to a defined group of services. By way of example: place name Albertinia, all hospitals and medical practices form a "spatial cluster of services". We then relate the number of people within distance concentric circles out from that cluster. What we have done in the past is to take approx 10 thousand claims - our development model runs about 100mil claims - and drill down on a parameter of the spatial relationship (age, sex, S/E status by salary, rank (for police & military), disease, claim pattern etc). We then sample from your dataset and do analyses from it to see if our spatial linkage provides a different insight to yours, ... and at first blush it seems good!"


Feel free to contact us, to discuss your requirements!



NEW! Dataset Extract.


SA Districts.

We have a "SA - districts" dataset available, as an extract from our postal dataset.

The list refers to the main towns (approx 450) in South Africa, within 1) provincial and 2) regional (garden route / bushveld / former Transkei / etc) parameters, and includes a gps waypoint associated with the area.

It is very powerful as a capturing tool - being used in the SA web-portal industry - as it allows 1) simplified capturing within 2) a searchable framework , offering both regional and provincial views of South Africa.

Rather than trying to manage 14 000 placenames, why not just use the main 450? 


Feel free to contact us, to discuss your requirements!


 _____________________________________ 



Old Telephone 

 

Kevod Insurance System

Our flagship solution for the intermediaries who manage their own schemes, collect premiums, issue policies and record claims.

Read More

Screen Shots

Screen ShotsTo take a closer look at how the system displays on your PC and the user interface.

Click Here

KIS Brochure

Brochure DownloadTo Download the latest KIS Brochure to read at your leisure.


Click Here

Copyright © Ks4 by Randmark 40 (Pty) Ltd

Aqua IT Consulting