In November, GBS co-director Harley Prowse attended the Esri 2018 GeoConX Conference in Dallas, Texas. Here he has provided some key take aways from the event, as well as some images from old sale yards in Fort Worth – including cowboys!
Overall, GeoConX was a good conference, with a focus on the use of the Esri platform for Gas, Electricity and Telecoms businesses. There were lots of user-based presentations along with the usual Esri technical streams. There was a big focus on Utility Network (UN) this year as you’d expect. It was also the first time that Jack Dangermond (Esri Owner and President) has been to GeoConX in person, and was incidentally the largest overall for attendance this year.
So what were my key take-aways from GeoConX?
- Some large utilities have been working with Esri since 2016 in order to be just now going live or about to go live with UN.
- The UN model for Gas is mostly ready (UPDM 2018) and a large east coast (USA) company has just gone live with their first region – the company is called NationalGrid, the project is the Gas Business Enablement. They will do electricity later.
- The current Esri Electricity UN data model is USA-focussed, so not ideal for the Europeans (and probably not ideal for NZ either). A few European companies have therefore got together and have developed their own Electricity Model and have made it freely available, asking for input to make it better. Senior Esri UN personnel have mentioned how fantastic it is, and have encouraged the US-based companies to look at it as well. You can get it and find out about it here. They look like they have done a really good job so far – focussed on electrical domain networks at this point, while structures are next.
Here were a few lessons learnt from other companies in attendance at GeoConX:
- Start early – this is a big change, focus on the data, understand your data and have a clear strategy.
- UN is way more fussy with data errors than Geometric Network (GN), so migrate early to find the issues. There’s no other way to do it really, and trying to write data inspectors to find the issue is a waste of time. They advised to just migrate a copy and see what happens, then fix in your existing system so when you come to migrate for real, the data is better. Starting early allows time to find and fix data issues.
- Make sure your 3rd party vendors are UN ready, or at least have a plan. There’s no point in going to UN if they are not ready, though you can still commence test migrations to work on your data.
- While you can do lots of cool stuff with UN (containment and high fidelity modelling), you don’t need to – and by keeping it simple initially and keeping the costs down, the cool stuff can come later.
- The way you integrate with other systems (services based) will be different to how you do it now, so the sooner you start the better.
- This is a chance to fix all the old issues you’ve always wished you had the opportunity to fix, so start early.
- Keep an eye on what you can’t control – such as 3rd party UN readiness and releases of UN (what’s new, what’s coming etc). You don’t want to get ahead of them.
- Sell it to management, not by how cool the UN is, but in business terms of cost, benefits and risk.
- Consider running a UN in parallel – a non editing version of UN – use it for reporting and integration and building new apps, keep doing the editing in the existing system. This allows you to build experience in the UN and get some new UN-based deliverables (reports, apps, integrations) up and running. It will also allow you to build up your ETL experience and refine the ETL scripts (don’t use Excel, you will go insane!), and fix data issues before the final migration to full UN. This can allow you to reduce ArcObjects dependencies – i.e. don’t build any new ArcObjects based stuff unless you have no choice.
- One presenter at GeoConX has been through paper to digital, migration from non-Esri (Smallworld I think) to Esri GN and now GN to UN – he said the GN to UN was definitely the most complex, so start early!
- Make sure you have some key and easily repeatable data validation steps in your ETL process:
- Record counts
- Data joins
- Date fields
- Invalid geometries (GN allows this, UN, not so much).
- Check your projections (not a major for NZ).
- Building the UN
- Extract data (start with a small sample)
- Build the network
- Profile the results
- Remediate (snapping issues, overlapping geometries, rule changes)
Benefits of going to UN:
- Implement Esri next generation software that will last 10-20 years – platform for growth
- Cross platform solution – similar functionality across desktop, web and mobile (improved data visibility = better decision making, efficiency)
- Advanced network model for GIS which provides a much better representation of how things are built in the real world
- Data quality improvements
- Services based solution – easier integration, more integration options, simplify integration = more robust, easier to maintain.
- Opportunity to streamline and improve related processes
- Improve compliance with regulators, improved safety.
- Improved customer service
- Replace ageing system to improve performance and reduce risk
- Deliver continuous improvement
- Can be fully cloud based if desired (this is what NationalGrid have done)
- Look for an opportunity to build the upgrade that many need to do (10.2.1 to 10.6) into the UN project for potential long-term savings – don’t 2 step it, do it once and go straight to UN now if possible. This may increase you current upgrade project costs, but will reduce costs as you won’t have to do a 2nd bite before 2024 when 10.6.1 support ends.
Mitigating Risk with UN project:
- Go Agile – well suited for migration projects where there is a need for many iterations and things are changing regularly – NationalGrid said they would not have done it any other way. Although it was a bit scary initially for management, it has allowed the delivery of continuous improvement.
- Engage UN expertise on the project, work closely with Esri/Distributor and Business Partners.
- Prioritise complex user stories early, as this allows time to deal with any issues and develop work arounds.
- Some functionality missing (e.g. dimensions and lack of line split capabilities) – so needed to assess if interim solutions could be developed. I think these two issues have or will soon be addressed by Esri, but it is still a good example of the approach to think about.
- Leverage change management expertise to manage expectations with the end users – it will be a big change for them – new tools and processes, new data model etc.
- Start early
So, if you had not already picked up on the main message at GeoConX, it’s start early!