The state of Washington is coordinating an ambitious effort to normalize transportation data from multiple counties into a common repository called the Washington State Transportation Framework (WA-Trans). The site will have a data user portal and a data provider portal and automated QA/QC methods that will include data change detection. V1 Magazine editor Matt Ball spoke with Tami Griffin, WA-Trans project manager, about the technical and collaborative hurdles that she is addressing in order to provide the best up-to-date data on roads, railways, ferries, aviation, ports, and non-motorized transportation infrastructure in her state. The state of Washington is coordinating an ambitious effort to normalize transportation data from multiple counties into a common repository called the Washington State Transportation Framework (WA-Trans). The site will have a data user portal and a data provider portal and automated QA/QC methods that will include data change detection. V1 Magazine editor Matt Ball spoke with Tami Griffin, WA-Trans project manager, about the technical and collaborative hurdles that she is addressing in order to provide the best up-to-date data on roads, railways, ferries, aviation, ports, and non-motorized transportation infrastructure in her state.

V1: From what I’ve seen, WA-Trans is a very advanced effort.

Griffin: I appreciate that. People I have talked to in other states aren’t doing anything  similar to what we are, and only a few are thinking of something even close to this.

V1: It’s very ambitious, particularly in terms of the amount of data providers to coordinate. Was there a foundation within your state for that level of collaboration?

Griffin: No [laughing]. There are organizations such as WAGIC (Washington Geographic Information Council) that encourage collaboration, but they don’t have funding mechanisms and other structures to support it.  I approach things differently because I am not a GIS person, I am an IT person with a public relations and communications background. So, when I was given this job, rather than looking at this as a technical exercise I looked at it as a business exercise.

I gathered as many different groups that are interested as I could, and then out of that we selected a steering committee. I continued to communicate with all the different groups, and the steering committee worked very hard with me. They are the ones who really helped me with the GIS end of things initially.

The first thing I wanted to do was a business case assessment and business needs assessment. The collaborative part of it came from recognizing the business needs we share, and our commonalities. We did this collaboratively right from the start, and I was very fortunate to have some really collaborative people who have committed a lot of time from the very beginning to work with me on this.

V1: Did it become apparent where the strengths and weaknesses were in terms of your data layers, and was data quality a big part of the business case?

Griffin: We didn’t really focus on that. We are running into more of that now as we implement. There definitely will be some areas where data are not going to be quite where we would like them to be at first.

The strategy has always been to bring data quality issues to the light of day, and then use information and positive feedback to create pressure for change and improvement. A lot of local governments are doing the very best they can, and they will do better if they are given an ability and a motivation to do better.

One of the things that we do is we produce reports and give them back to the data providers about their data, and it tells them things that can be improved, but we don’t make judgements about what we see. We just tell them things like, “you have this network that is not connected to other networks.” Or, “your linear reference system has these characteristics, some of which may be anomalies.” And then they can decide if those are things they need to fix. A lot of times they’re really thankful, because they just don’t have that information.

V1: How has this effort been funded?

Griffin: I’ve had to do everything with grants. I managed to raise a fair amount of funds, but we are getting to a point where we have to raise more funds within the next year or so. My position and another position has been funded by WSDOT and we have partnered with other state DOTs who have helped with funding through a pooled fund mechanism.  Despite that fact we are making a lot of progress.

Our goal is that by June 2010 we should have sixteen counties completely done (out of 39), plus we should have 75 percent of the infrastructure completed and ready to promote to a production environment. The infrastructure is in a test environment now in terms of maintaining the data, allowing online data update, transforming data, and then running them through various back end processes to maintain the data. We’re also working on a data user portal to serve the data.

V1: Are these first 16 counties the most populous, and were some of these counties chosen because they had a good start on data collection?

Griffin: We do have funding specifically to focus on the most populous counties, but we also have other funding to focus on some of the other counties. So, it’s a mix.

The first few were chosen because they were very sophisticated and they had a lot of knowledge to help us to figure out how to do this. They were also partners and they wanted to be involved very early on.

The next few were chosen because they had fairly good data, although maybe not as sophisticated systems, but they also had long-term data sharing relationships with each other. So they were very comfortable with the idea of sharing data. They were also in the central eastern part of the state, so that meant we weren’t totally west coast focused.

And then, of the other two, one of them is being driven by business needs to be able to locate collisions accurately, and then we have some funding through a U.S. Geological Survey grant to partner with Spokane County, which is in the eastern part of the state.

V1: What percentage of the project would you say is focused on the data infrastructure?

Griffin: I would say 45 to 60 percent. It will change overtime as we get the infrastructure farther along and focus more solely on the data.  We have a full-time contractor, and he is very good. And then we have a technical lead that spends probably 50 percent of his time on data and 50 percent of his time on the software end of things.

Then we have a GIS analyst and myself, and I probably divide my time between data and between fundraising and public relations type stuff, communication and project management, all of that is part of my role. We have a relationship with Central Washington University, and they do a lot of work for some of the data.

V1: In reading over the WA-Trans Framework on the web site, it indicates that the overall mission is to provide multi-modal transportation data. To date, are you working on the road data?

Griffin: The road data are provided by counties generally, and that’s the initial goal because it’s the low hanging fruit and has the greatest overall benefit. Then we are adding some DOT road data, linear referencing systems and address ranges.

V1: We’ve talked primarily about data sharing. Are there plans to work on applications that use this data to create map products or to output this information in report form for specific purposes?

Griffin: We are not doing much of that yet, although we will be. We are getting some funding specifically for reports for a collision location system.

We have a data user portal that has been developed and is being tested right now. It allows extraction of data by county boundary or by data provider or by a polygon, or by the whole state. It allows some online data viewing and simple query. The ultimate goal will be to have various products coming out of that, for instance a product for address geocoding, a product for transportation planning, a product for various referencing systems. But initially, these capabilities are very limited, because we need to start figuring out how to build some of these products, and they’re all funding dependent, but it’s all part of our vision.

V1: How does the data provider portal work to ingest data from the counties?

Griffin: The data provider portal is set up to deal with known entities that we have relationships with. We need to know enough about their data to build them a transformation process that’s unique to their data, and that transformation process is handled in Safe Software’s FME Server that links to the portal. Providers have the ability to log into the portal and to upload data, and then it is transformed to our schema and our projection and put into a loading database. From there the data is evaluated, various tests are run on it, including change detection against the previous data that they sent us so that only the changed data is moved forward. Also, other QA/QC things are run directly on the data.

This is where we give the feedback. We may find something that’s fatal enough that we have to have them fix it before we can go any further. But, the assumption is that their data is their data and we are not going to make judgements about data quality. And we are also not going to refuse data unless it doesn’t meet our minimum standards.

We have to have a minimum QA/QC for our processes to store properly, but our goal really is not to spend a lot of time evaluating other people’s data, it’s to ingest it, to integrate it, and then to make it available for use. Once it has gone through the change process and this initial QA/QC, then the changed data is promoted into what we are calling the stating or editing environment. And that environment involves integration with data around it, it involves structuring the data appropriately, and dealing with any linear referencing materials, and then from there the whole data is promoted into the production database and is available through the data user portal.

V1: You mentioned that one of the business drivers is collision data. Are there federal mandates for that part of this service?

Griffin: There are federal mandates, but it’s more than that. It’s really how the transportation dollars are distributed and  anecdotal information being used for those decisions or can you really relate incidents of fatal and disabling collisions to roadway metrics and other characteristics on the roadway. Who knows, maybe it’s not anything to do with the road, maybe there is a bar down the road, and we have drunk driving collisions. It’s about finding the best solution, and the best use of the public funds to solve the problem. It’s not always just laying down pavement.

The goal is to eventually be able to use collision data on both the local roads and the state roads as accurately placed as possible in a geospatial environment combined with a lot of other information to do a root cause analysis in order to spend the tax dollars, the federal dollars and the state dollars, in the most appropriate way for safety mitigation.

To date, the state sometimes has had difficulty locating collisions with the best accuracy based on collsion reports and they are trying to figure out areas where there are the most collisions.. They also don’t have the ability to look at that data spatially in relationship to a lot of other things because it comes in as a description that is related to a whole bunch of other location information, that isn’t always using the same location referencing system.

A standard underlying base map is really the issue that we are involved in. That’s something that addresses the whole complexity of the problem because if there isn’t a standard underlying base map we can’t support all these different queries. Collisions are only one business need that we happen to be focused on now because we have a specific request. We have many others at different levels of government.

V1: When you did the business case analysis was there also a cost benefit analysis?

Griffin: I was supported by the Federal Geographic Data Committee to work with the Geospatial Information Technology Association to do a cost benefit analysis. From that, we know that over twenty years there is a minimum of twenty-six millions dollars of benefit across the state. But that’s the tip of the iceberg, because I was only able to talk to four state agencies and six counties, because we had to do it within a fairly short amount of time.

The benefits that have been realized right now are the work that we are doing with counties on their data. Counties are taking the data back and using it themselves in its improved format and they are finding value in that. And I think in the collaboration the relationship building that is going on has been very beneficial.

We’ve also set up a pooled funds study that has really been helpful because we have other states that want to do something like this and they are each giving us a small amount of funds, and we are sharing the results with them so that they can use it in their own environment.

V1: What are some of the next steps that you’ll take when you have additional funds?

Griffin: We are looking for more states to be involved in the pooled fund so that we can complete the software work we have identified in our test environment. And then we will be looking for funding to promote all of that to a production environment.

We are going to be also looking for funding to complete the rest of the state and other modes, and then eventually we would like to have some funding that we can use to help the locals maintain their data. I think we are going to find that really the best bang for the buck is letting the people who create the data improve the data, giving them motivation to improve the data and assisting them with identifying areas where there are problems. The collaborative approach is best, because then it benefits everybody in the state instead of just benefiting one particular group or another.

V1: There is call right now for a national GIS. How can your approach influence such a plan?

Griffin: I have been involved in Transportation for the Nation, which is an initiative of the National States Geographic Information Council (NSGIC). They looked at what we are doing and they said, wow, you are doing what we want to see everybody doing and you are actually doing it successfully in partnership with other states.  We have also provided data to The National Map and will continue to do so if they are interested.

We have done some cross border work with Oregon, and we are looking at maybe doing some with Idaho, and eventually we would like to do it with Canada. Our pooled funds study has different states involved, so they are also doing cross border work as a result of this with some of the other states that are involved. I think that’s why we have gotten a lot of attention.

I think the answer is to get data from the people who have the most knowledge about the data values, the correct data values for addresses, for linear referencing, for roadway metrics. And the reason this data is best is because they have to use it for their own business processes on a daily basis.

You will never be able to purchase data that has the temporal accuracy and has the same commitment as local data. I think in the long run we’re going to end up being a cheaper alternative than licensed data. It may take a while before that happens, but in the long run we will be, because we don’t have to deal with the licensing issues and you’re not at the mercy of vendors who’s business needs are not the same as the government

< Prev		Next >