Gas Explosion Pinned On Poor Data

Posted - Hayley Merrill:  You may recall the news headlines from September last year in which a Pacific Gas and Electric Co (PG&E Corp.) pipeline exploded in the San Francisco suburb of San Bruno, California, killing four people and destroying 38 homes.  The hearings have just recently begun and are examining the circumstances that led to the fatal explosion. It has become immediately apparent that the company is missing critical information that keeps track of its pipelines including location-intelligence, potential failure points and maintenance records.  In fact the company has stated that is has had a real challenge converting paper records to computer files for a number of years now. 

A San Francisco Chronicle article suggests that software upgrades and errors in managing the utility’s information system were contributing factors to lack of maintenance of the pipeline.  It appears this is not only a significant challenge for PGE&E Corporation but many other utility companies worldwide who are finding the management of geospatial information really complex and that the value behind its intelligence is grossly underestimated.  Mary Muse, a PG&E Corp. gas engineer, involved in the computerisation of old pipeline paper maps to digitised maps, told a convention of mapping experts, “Validation of individual field values was not performed.  This led to incorrect or inconsistent values being populated in the fields." 

Whether an organisation manages the capture and validation process itself or outsources it is critical that good (and known) data quality underpins business planning and decision making.  Surely it has never been more critical to recognise the valuable asset that location-data is – especially if it helps mitigate disasters like this.  

Poor Data Quality a Crime

Posted - Bob Chell:  As we decided to launch a blog for 1Spatial Consultants, there is no better current example that brings focus on data to the people’s minds than the Police Crime Maps.  My honest first thoughts, when I heard the site had been down due to demand, echo’s that from Adrian Short’s blog - ‘too much traffic to our website is a problem we’d all like to have’.

As I got the train from London to Cambridge, I then saw that what we could also get hold of was the raw data itself – fantastic!  Then, just as we were arriving into Cambridge, I actually started to read numerous blogs about issues people actually had with it all. All these have already been written about and I found informative summaries from Steven Feldman and the previous mentioned post from Craig Short to cover everything.

My attention is always drawn to comments about ‘understanding about the data’. One of the first articles I came across was on the Guardian website, which contained an eye-opening ranked list of a streets and crimes.  Interestingly, Cambridge had a hot spot, Peas Hill, in at number eight. Working and living in Cambridge means I already have context about this place - I know this area. And I also know that Jamie Oliver recently opened Jamie’s Italian which is right in the same place, right on the corner of this street - and it’s not a particularly long street! I’m pretty sure that Jamie Oliver was not advised to open a high profile restaurant in a crime hot spot. So I became a little more curious about this information. What rules were there behind its creation?

Having got a basic glimpse of the data, I needed to investigate further since the raw data themselves without context have limited value. So Step 1 - discover and learn more about this data’s provenance and governance. Without this type of understanding, you won’t really be able to make an informed decision on whether the data is suitable for you or your business. Is this suitable for you the mission that you and your company are on?

I share many people’s views that this is huge improvement on what we had before. I’m also slightly biased, having worked with the likes of British Transport Police, and empathise with the effort that they are putting into making continued improvements in managing the vast quantities of information that keeps passing by their desks.

The data are now open. However we are simply at the beginning. First we need a baseline to work out a where we are with the quality of the data, then we can start to determine how to improve it to make it fit for purpose, give it context. There appears to be valid thought-processes or rules that have gone on to justify why the information is as it is. So all this needs to be put into context, so that the issues and understanding around inaccuracies are not what these transparency initiatives are remembered for.
The data are available, so the government is hoping that others can easily look to take a different approach to using it and analysing and presenting it in novel ways. I like the ITO world updates to Open Street Map as Berners-Lee showed at TED2010.

There are other examples already appearing, I’m sure you can find them through the search engines. One thing is for sure though, the now monthly snapshots of crime figures for England and Wales should plug straight into all sorts of Business Intelligence tools and I’m looking forward to how this grows up. I might look at Oracle Business Intelligence Enterprise Edition (OBIEE) and others in the Consulting group will no doubt look at BI tools of their own preference. Keep an eye out for future posts.

Data Improvement Blog

Posted - Bob Chell:  I have been consulting at 1Spatial for over 10 years, and have worked with some exceptional colleagues and great clients. We have spent almost all our time solving the difficult, data-intensive problems in the spatial arena, problems that many people do not know how to solve. And along the way, collectively we have picked up a vast range of valuable skills and experiences.

Like the company we work for, despite changes in focus over the years from the GI industry (desktop, web, centralised databases, web-services) we have always maintained a constant focus on  your data - ensuring its quality and looking to automate as much of the data quality management and improvement process as we can.

We are way overdue in sharing some of our experiences, so this is our first foray into rectifying that, so I hope you find the posts interesting.

I plan to use this blog to allow the whole team to keep people up-to-date with the projects and discoveries that we all make along the way. And as this is the first post, I thought you might like to know about the types of things we are currently working on.

Matt Beare co-chairs the OGC Data Quality and is in the closing stages of ESDIN Project, where we have been contributing to a number of successful work packages. Although based in our Cambridge office, this has kept him busy right across Europe.

Tom Spencer supports PSMA on site down in Canberra, Australia. He also has a trip out to Land Information New Zealand (LINZ) in the near future to get them up and running with Radius Studio, as well as supporting our Managing Director for Asia Pacific Guy Perkins.

Chris Wright is also in our Consulting and brings over 20 years of experience to the team. He recently helped establish our Australian office, (now operated by Tom), originally helping MidCoast Water using Radius Studio to produce Geo-Schematic data from MidCoast's TopoBase database and engineering data quality business rules during the migration to TopoBase from Munsys GIS.

From Belgium, Luc Van Linden has his time divided between looking at our ventures with INSPIRE and helping Tom with some project work for Queensland Transport and Main Roads (QTMR). He has also recently found some time to look at Deegree OSGeo project, and he keeps his number 3 status on the Oracle Technology Forum too.

I'm currently working with English Heritage to assist them in future strategic guidance around interoperability. I’m also working on data migration and quality improvement programmes at the Environment Agency. Outside this I have also started to implement solutions using GeoServer and explore OBIEE 11g.

As you can see, it is a talented team that covers a range of tools and technologies. So if you want to find out anymore, get in touch.  You can email us at: consulting@1spatial.com.