bridging the gap
DESCRIPTION
Bridging the gap between relational and spatial data How data quality links customer to spatial data sets see http://www.masterdata.co.za/index.php/geocoding-cresTRANSCRIPT
Optimising the value of your information asset
BRIDGING THE GAP
Gary AllemannMaster Data Management
Optimising the value of your information asset
Different ways…
Optimising the value of your information asset
...to represent location…
Optimising the value of your information asset
..means mistakes happen!
Optimising the value of your information asset
Ways of representing location
• Descriptions – unstructured text description• I live on Piet’s farm next to the mielie
fields.• Take the first right after the petrol station
and we are the white building on the left.
• Images
Optimising the value of your information asset
Ways of representing location
• Address – (Semi) Structured used to represent the location of a building or property
• 14 Fifth Ave, Rand Park Ridge, Jhb• 5delaan 14
Randparkrif2156
• Corner of Fifth Avenue and Rand Road, Randpark Ridge
Optimising the value of your information asset
Ways of representing location
• Geocordinates – a precise location on a map
May represent objects that do not have an address
• Meter• Road• ATM
Optimising the value of your information asset
The Focus RoomsCnr Kikuyu &
Leeukop StreetsSunninghill2157
Relational vs Spatial
Which corner?
Optimising the value of your information asset
The Focus RoomsCnr Kikuyu &
Leeukop StreetsSunninghill2157
Relational vs Spatial
Spatial represents an area
Optimising the value of your information asset
The Focus RoomsCnr Kikuyu &
Leeukop StreetsSunninghill2157
Relational vs Spatial
-26.035557,28.065022We select a point to represent this address
Optimising the value of your information asset
The Focus RoomsCnr Kikuyu &
Leeukop StreetsSunninghill2157
Relational vs Spatial
-26.035774,28.064682May vary (slightly) from one reference set to another
Optimising the value of your information asset
Typical South African address issues
Address 1 Address 2 Address 3
44 Gleneagles Road Greenside 2199
Tweedelaan 48 Nelville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktwon 2193
Aberdeenstraat 122 Melville 2092
101 Greenway Greenside
Stores Pool Bar Main Road 2092
Gleneaglesweg 42 Greensde Jhb 2034
No address standard – mis-fielded data
Optimising the value of your information asset
Typical South African address issues
Address 1 Address 2 Address 3
44 Gleneagles Road Greenside 2199
Tweedelaan 48 Nelville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktwon 2193
Aberdeenstraat 122 Melville 2092
101 Greenway Greenside
Stores Pool Bar Main Road 2092
Gleneaglesweg 42 Greensde Jhb 2034
English and Afrikaans address data
Optimising the value of your information asset
Typical South African address issues
Address 1 Address 2 Address 3
44 Gleneagles Road Greenside 2199
Tweedelaan 48 Nelville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktwon 2193
Aberdeenstraat 122 Melville 2092
101 Greenway Greenside
Stores Pool Bar Main Road 2092
Gleneaglesweg 42 Greensde Jhb 2034
Abbreviations, Spelling and Typing errors
Optimising the value of your information asset
Typical South African address issues
Address 1 Address 2 Address 3
44 Gleneagles Road Greenside 2199
Tweedelaan 48 Nelville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktwon 2193
Aberdeenstraat 122 Melville 2092
101 Greenway Greenside
Stores Pool Bar Main Road 2092
Gleneaglesweg 42 Greensde Jhb 2034
Missing information or wrong post code
Optimising the value of your information asset
We need a good quality address
Address 1 Address 2 Address 3
44 Gleneagles Road Greenside 2199
Tweedelaan 48 Nelville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktwon 2193
Aberdeenstraat 122 Melville 2092
101 Greenway Greenside
Stores Pool Bar Main Road 2092
Gleneaglesweg 42 Greensde Jhb 2034
Missing information or wrong post code
Optimising the value of your information asset
Semi-structured to Structured
Number Street Name Street Type Suburb CityPost Code
44 Gleneagles RoadGreenside 2199
48 Tweede laanNelville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktwon 2193
122 Aberdeen straat Melville 2092
101 Greenway Greenside
Main Road 2092
42 Gleneagles weg Greensde Jhb 2034
Optimising the value of your information asset
Standardised & fix (common) errors
Number Street Name Street Type Suburb CityPost Code
44 Gleneagles Road Greenside2199
48 2nd Avenue Melville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktown 2193
122 Aberdeen Street Melville 2092
101 Greenway Greenside
Main Road 2092
42 Gleneagles Road Greenside Johannesburg 2034
Optimising the value of your information asset
Enrich – add / correct missing info
Number Street Name Street Type Suburb CityPost Code
44 Gleneagles Road Greenside Johannesburg 2199
48 2nd Avenue Melville Johannesburg
18 Park Lane Parktown Johannesburg 2193
19 Park Lane Parktown 2193
122 Aberdeen Street Melville 2092
101 Greenway Greenside
Main Road 2092
42 Gleneagles Road Greenside Johannesburg 2199
Optimising the value of your information asset
• International Sources• GPS, Mapping
• Local Sources• Surveyor General• Local Government• Commercial Sources
• Each has strengths and weaknesses based on your requirement and location – test to fit your need
• For example• 14.5 of 52 million South Africans live under
Trad Authority
Compare to reference data source
Optimising the value of your information asset
• Delivery and collections• Plan routes based on proximity• Identify and resolve errors before sending
driver out
• Territory management• Assign clients to appropriate reps
Improved planning improves efficiency
Optimising the value of your information asset
Case Study 1: Route Planning
48 2ND Avenue
101 Greenway
44 Gleneagles Road
18 Park Lane
Optimising the value of your information asset
Case Study: Route Planning
48 2ND Avenue
101 Greenway
44 Gleneagles Road
18 Park Lane
19 Park Lane does not exist in reference set!• Could have been wasted trip• Can either:
• Assume next to number 18• Call and confirm address before
travelling
Optimising the value of your information asset
• Planning based on 2001 Census data• Current population levels assumed
• +- 100million address records geocoded
• Geocoding gave• Understanding of area dynamics• Spatial targeting for services• Identification of delivery bottlenecks and
disparities
Case Study 2: Delivery of services
Optimising the value of your information asset
Standardise e.g. +- 600 variations of East London
Optimising the value of your information asset
Case Study: Delivery of services
Optimising the value of your information asset
Actual beneficiaries vs Assumed
Implications• Some areas underserviced e.g.
Umtata• Some areas either have an over-
allocation of resources e.g. Port Elizabeth or there are many candidates for services that are not registered in the area.
Optimising the value of your information asset
• Bridging the gap between address data and spatial data can bring significant benefits
• Different applications require varying levels of accuracy
• Data cleansing brings the improvements for address accuracy necessary to bridge this gap
Conclusion
Optimising the value of your information asset
• Gary Allemann• +27 83 632 1591• [email protected]• www.masterdata.co.za
Questions