Go to content Go to navigation and search

Home

Straws in the Wind Blog Articles


Search

Browse

RSS / Atom

Email me

textpattern

Creative Commons License
All Blog Articles, Data Models and Free Source Code by Simon Greener, The SpatialDB Advisor is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

Talk on Open GeoData in Australia

Wednesday January 21 2009 at 00:30

Yesterday, Tuesday 20th January 2009, I gave a flawed, imperfect, rather long talk called Free And Open GeoData – From Shadows to Reality at the March South Together, linux.conf.au 2009 Conference at the University of Tasmania.

If one wants to get a message across, and perhaps be entertaining, one has to sprinkle one’s talk with deliberately harsh criticism of the status quo.

The simple reality is that access by individual people to large scale (accurate) government GeoData in Australia (principly road centrelines and street address points) such that they can value-add, remix or re-distribute their creations, is way too expensive and restrictive. Other government departments, private businesses and Google – via PSMA – are well catered for by the centralised state government departments that integrate local council data into single, high detail, state-wide datasets. So, saying and trying to explain why the situation exists for individual people, by trying to describe the Eco-System that public servant geospatial producers live in, will include inflammatory statements.

But in my talk I am also critical of the OpenStreetMap, Track2Australia etc projects even though they have all the right intentions. I think it is more important that we create a project for the capture of simple street address points for the purpose of free, online, geocoding.

I also make the point that “channel marketing” (if that is the right term) which sees hardware vendors purchase or team with software (data) vendors to create bundled products is both a limit to our freedom but may also, inadvertantly, be a key to unlocking government data in the future.

If I have gotten any facts wrong, please email me and, if you convince me that I am wrong, I will modify the talk on this website accordingly

Creative Commons License

post this at del.icio.uspost this at Diggpost this at Technoratipost this at Redditpost this at Farkpost this at Yahoo! my webpost this at Windows Livepost this at Google Bookmarkspost this to Twitter

Comment [4]

Access to government datasets is becoming increasingly frustrating, especially when you work for another level of government! We are consistantly being approached by Government Departments (both State & Federal) to improve their own datasets – in the main we dont have concerns regarding this (especially when it comes to data like Road Centrelines), but Im noticing a trend towards the use of the term data-sharing but none have real intentions of sharing back datasets which we have contributed to.

Google Maps causes us headaches on a monthly basis. There is a serious lack of information about where data is derived from and it is Council that angry enquirers get referred to in order to make the issue go away. I dont know in our area where PSMA are getting their data from, because its not correct.

Addressing is quickly becoming a serious issue – most major companies now are relying on some semblance of a GNAF, but I cannot get a definitive answer where from (I presume PSMA). We get calls all the time saying that addresses are not accepted by many companies because they don’t exist in their GNAF – Council’s are the derivation authority for addressing, but yet again we get overlooked when it comes to updates.

The lack of a strong Spatial Strategy that covers all levels of Government will continue to hamper the access of data. There needs to be a shift away from having each level of government deriving their own data on the same themes (I accept there is a need to derive at a higher level of state/nation datasets, but these should be fed from the ground up).

I agree that data capture is expensive and it is this sheer expense that drives the need to recover some cost involved the capture, maintenance and upkeep. But, this should not be an excuse to lock data away or into expensive and / or restrictive uses.

— LocalMan · 21 January 2009, 07:29 · #

LocalMan,

You raise some interesting issues which reinforce what I presented as being the situation in some states.

So, when the state government departments come to your council for improved data (at least they do so), are you saying it is all one-way with respect to data-sharing?

If so, this is quite poor.

I can understand about Google Maps after all that “©2009 MapData Sciences Pty Ltd, PSMA” on the map says it all. MapData Sciences are the PSMA data reseller that provides Google with Australian road centreline (though I would have thought the PSMA road centreline data was nowhere near as accurate as Navteq for navigation), parcel and address point (GNAF) data. Of course I have no idea what the arrangment is between Google and its Australian supplier, but not having the local councils in the loop seems to me to reinforce what I am saying about complicated supply-chains and market-focus. Didn’t they consider that wrong data on Google would raise complaints at the council level and so arrange something? My experience as a GIS Manager was that the local state agency responsible for state-wide datasets were neither proactive nor seemed to have healthy relationships with all its suppliers. It may have changed – but I would want to hear from all suppliers and not just the main integrator!

I am surprised that a council doesn’t understand that G-NAF is Google’s source of address point data. The lengthy update cycle for errors in G-NAF are real because to get it right they really have to talk to all the councils (via the State agencies!). Note that the PSMA cannot, when constructing G-NAF correct obvious errors in the data: this has to be sent back to the relevant state and then back to the source council. This is complicated and appears not to be done as well as they would like.

With regards your point about councils being the “derivation authority”, I would argue that a council property address is not the only source of valid addresses. For example, Australia Post’s Names Address File (NAF) is larger, more accurate and comprehensive than the PSMA’s G-NAF and included non-Property addresses (which Police, Fire and Ambulance use). Yet will these ever appear in Google?

I agree about Spatial Strategies but it would appear that so much paper has been expended by the spatial industry in Australia over the last 15 years on spatial data, metadata, discovery, distribution, aggregation and licensing that they could have claimed Green Credits for years if they hadn’t written them!

There are many models for organising government departments and data. I like the UK’s 2-tier system but it can’t work here in our 3-tier system. I don’t know anyone who likes the geodata licensing of the Ordnance Survey!

I agree with your last paragraph. The lock up and cost seems to me to be linked to the heavy involvement of government departments. As I said in my talk, government departments are not known for their flexibility, efficient and effective business processes and market focus.

But we have to find some way through it all.

Simon Greener · 26 January 2009, 07:24 · #

Simon,

So, when the state government departments come to your council for improved data (at least they do so), are you saying it is all one-way with respect to data-sharing?

Normally, councils have a data sharing agreement with Department of Lands (DoL) for their DCDB and work closely with them on improving cadastral quality in their LGA. Most councils welcome this because for a long time there has been a divide between councils and the DoL on a number of levels. Sadly councils’ hands with this data sharing has been forced by a separate government department (Department of Planning) who, without consultation with either councils or the DoL, have released guidelines for the production of Local Environmental Plans based on the DCDB held by the DoL. Classically, the left hand isn’t talking to the right hand!

On the opposite side of this, the Emergency Information Coordination Unit (EICU – housed within DoL), shoved a data sharing agreement with a 2 page list of GIS information that they were looking for at councils. In return they get their own data back in a standard format across the state for their Emergency District (ED) along with data from other councils in our ED. I can understand why they are being asked for it, but the manner in which it has been asked for and what is being offered in return doesn’t make it “data sharing”. It’s really data giving (I don’t understand how a re-purposing of the data being provide can be considered sharing!). What makes matters worse (especially for council finance personnel) is that EICU are asking for councils’ most valuable GIS asset, their Aerial Imagery (providing it legally isn’t a problem, because councils own the imagery outright), without charge!

Some departments we have data shared with provide data that is not appropriate to a council as well – this appears to be an unfortunate evil due to the fact that councils fall inside the auspices of the coastal regional offices.

Didn’t they consider that wrong data on Google would raise complaints at the council level and so arrange something?

Google’s response to this has been somewhat of indifference – my experience has been that Google will brush your concerns off to a range of agencies, some of which either are not aware that they are involved or are simply not involved.

What has become worrying is that while the Cadastral data (and geocoding data) has moved on, the road data has not which means that the effectiveness of the service is diminishing.

I am surprised that a council doesn’t understand that G-NAF is Google’s source of address point data. The lengthy update cycle for errors in G-NAF are real because to get it right they really have to talk to all the councils (via the State agencies!). Note that the PSMA cannot, when constructing G-NAF correct obvious errors in the data: this has to be sent back to the relevant state and then back to the source council. This is complicated and appears not to be done as well as they would like.

You might be surprised but there is very little knowledge at a local government level of the existence of a G-NAF even from GIS staff. I put some of this down to that because councils do not interact with PSMA directly (as a whole) there is no real need to know. I know we are not being issued corrections that have been discovered in the G-NAF, I can only assume that they are being sent to DoL and either being corrected there with no data-flow backward or being ignored.

With regards your point about councils being the “derivation authority”, I would argue that a council property address is not the only source of valid addresses. For example, Australia Post’s Names Address File (NAF) is larger, more accurate and comprehensive than the PSMA’s G-NAF and included non-Property addresses (which Police, Fire and Ambulance use). Yet will these ever appear in Google?

I understand on this one there are moves afoot that will see Australia Post make some changes to their NAF – what they are I’m not sure, but I’m supposed to be meeting with Australia Post to discuss a number of issues that have arisen locally.

There are many models for organising government departments and data. I like the UK’s 2-tier system but it can’t work here in our 3-tier system. I don’t know anyone who likes the geodata licensing of the Ordnance Survey!

I honestly don’t think it needs to be rocket science – each level should be maintaining the data for the service they provide or responsibility they hold. (excluding Defence / Secret Squirrel). A state roads dataset (as an example) should be a composite of the Local datasets and the national a composite of the states (appropriately filtered). This should not be hard to achieve and can be made easier by widely accepted data structures at all levels – why should each agency “re-invent the wheel” when it comes to recording core attributes against a road (name, seal type, width, length etc) or a water pipe (installation date, diameter, material etc). We should have core dataset requirements and a “localised” requirement, so if Tamworth Regional wants to record say a localised road number on a road we can.

Inter-governmental data sharing should be encouraged and fostered, all the current situation achieves is Siloing of data.

LocalMan

— LocalMan · 4 February 2009, 02:57 · #

This article might be of interest to readers about how silly the Ordnance Survey of Great Britain (OSGB) can be (a reputation it has had for many, many years) because it is very appropriate to this discussion on OpenGeodata or more open access to local, state and federal government data.

Simon

Simon Greener · 4 February 2009, 03:27 · #