Visual-tm with Erica is a software system for holding transport origin-destination (o-d) data in a consistent database and statistically merging it, using DfTs ERICA5, to build trip matrices, which take account of double counting, multiple screenline crossings and the problems associated with motorways. The o-d database can include data from roadside, household, on-train or on-bus interviews and pre-existing matrices. This allows the o-d matrix to be built-up over time as more surveys are done and as their data can be incorporated into the database so as to give a progressively more precise account of the pattern of travel. So for example a County could start with a few roadside interviews (rsi) in one part of the County and gradually infill into the other areas as more data is collected over time, to eventually cover the whole county.
ERICA5 is the matrix builder and our Visual-tm software is its front-end. Visual-tm provides a user friendly interface for the ERICA5 control files as well as essential functions for cleaning the raw data, coding postcodes to geo-codes and geo-codes to zones, expanding roadside interview data to classified counts, tabulating and analysing the o-d database. The Visual-tm with ERICA5 combination maintains the o-d data’s internal consistency so that extracts, tabulations and analyses can be undertaken across the whole o-d database or for selections from it. The matrices it produces can be input into a transport model and used for scheme appraisal with TUBA. It plays an important role in helping to keep consistency in the appraisal of different schemes and in the appraisal of the same scheme over time.
Visual-tm with ERICA5 keeps your important o-d data in a safe database, as an essential resource for the future. It contains the tools you need to use it for developing your Local Transport Plan, for modelling your new schemes and for scheme appraisal. In Best Value, it is useful for setting targets, monitoring progress and demonstrating that they are being achieved.
Visual-tm with Erica holds a multi modal database from which models can be derived automatically. This loses none of the functionality of the conventional approach but provides some important advantages as follows:
- The database can be updated as new data becomes available and the models are updated automatically.
- The model functionality can be extended as more becomes known about the transport market segments
- The geographic area covered by the model can be extended as and when new data is collected
- The model can be made more detailed as required rather than having to build it into the model from the start
- Model detail is therefore developed for the places it is needed most rather than having to waste resources building a model for things that are not going to be needed – they can be built when they are needed.
- The o-d matrix can be in-filled as more o-d data is collected
- Model elements are directly traceable back to their source data so errors can be corrected and model improvements made
- Model validation is much easier because flows can be traced back to the interviews from which they were derived
- Retains all the functionality of conventional modelling
- Is much quicker and cheaper than using conventional methods
- It also builds better models because the whole process is transparent
- The database can be set up to derive bespoke models for bespoke purposes with their own zone system and network.
- Avoids concentration of knowledge on key individuals making maintenance, distribution, support easier
- Keeps valuable data as a key resource for the future
- Enables data, models and forecasts to be used for monitoring.
- Monitoring data is then available for future transport planning
- Models are then independent of software vendor
- Databases can be held in simple spreadsheets so that they are easy to use directly
The database is set up initially with whatever o-d data is available which can include new roadside interviews (RSI), household interviews, on bus/ train etc interviews. This o-d data is cleaned and expanded in the normal way and put into the database, which is set up so as to build the o-d matrices automatically. New data can be spliced-in by amending the matrix build specification. Matrices can be initially built for 24 hours all-purposes and vehicles for validation and they can subsequently be built by purpose, time period, vehicle type etc thereby simplifying the modelling process and is one way the database supports the progressive refinement of the model.
Further refinement of the market segmentation can be achieved by collecting new data with say household interviews so that matrices can progressively exploit a more refined market segmentation (e.g. car trips by type of parking, parking duration, company reimbursement policy etc.)
Transport models of an area are often built for one purpose but having served this purpose they are kept and used for other purposes. Some authorities have kept their models for a long time, sometimes updating them from time to time. Just keeping a model is expensive and updating it is more expensive. On the other hand models can be use for detailed studies, where perhaps new data is collected across a new screenline to augment the matrix. The authority gives their model (or parts of it) for the study and the new data is put into their version of the model, used for the study and then discarded at the end of the study. This is a waste of valuable resources.
In Visual-tm with Erica, the modelling database would be to give to the study team, who insert the new o-d data into the database and derive their own matrix from the whole database. They should code their new infrastructure into the database and derive their own networks. The also code their planning data and other inputs. Having completed the study, the complete database including the new o-d data and new coded infrastructure links would then be handed back to the authority having been updated by the study team. The model improvements would then be available for everyone else so that someone doing a scheme in one area can therefore use the coding for all the other schemes without having to recode them all again. This approach helps keep the model up to date. It also makes it more detailed where the detail is needed and leaving it coarse where the detail is not needed. If the study is investigating new modes of transport and they undertake stated preference surveys this data is also put into the database and made available for future studies so that a light rail study in one area can use coefficients (e.g. the value of time) derived from another area – at least initially.
Hitherto model builders, having built their model throw away their data. This is a waste. Data is a valuable resource that should be retained for the future so that it is available for use over and over again. Many times have we had to use models that can’t be extended or adapted because the original data has been lost! It is a valuable resource, which should be kept for future generations. Keeping the original data can help plan the data update cycle. It can reduce the need to collect new data. Even when data has been replaced with more up to date data there is often the need to go back and investigate changes over time or to back-project the models to ensure temporal stability, or to see whether for example mode shares have changed over time. The modelling database ensures that the data – even historic data - is available for future users.
The data needs to be held with all the information needed to understand it (it’s meta data) as well as the objects, classes and methods used to manipulate it. These should be kept with the data so that the database retains all the functional processing needed to understand, manipulate and use it - held in the right way so that users do not have to reinvent the wheel every time they use the data.
The database can hold data as follows:
- origin-destination (o-d) data: roadside, household, bus, train, air, passengers and freight data.
- transport infrastructure networks: road, rail, bus, train, walk, cycle, air, passengers, vehicles and freight
- counts: vehicles, occupants, classified, automatic, number plate surveys etc
- speed, congestion, air quality etc
- current and forecast planning, demographic and socio-economic data
- forecast scenario networks
- forecast ridership, vehicle flows and revenue
The components and data can be chosen to suit a particular study. The zone system can be refined for a particular study by simply digitising the new zones, using the grid referenced origins and destinations to recode the o-d data to the new zone system and build study- specific matrices. (This relies on coding the o-d addresses on the trip records to grid references in the original data.) Specific networks can also be produced. The whole database can be put onto a geographic information system so that data can be displayed on maps, which can be easily cut-and-pasted into documents.
Visual-tm with ERICA comes as Visual-tm Corporate edition which includes all the functionality of Visual-tm plus all the functionality of ERICA5 and as Visual-tm ERICA edition which comprises Visual-tm with a 50-zone limit and no modelling capability and ERICA5. It is now available for PC’s running WINDOWS 2000/XP or above. ERICA is owned by DfT and contains Peter Davidson’s fast low-level object library for handling very large datasets and very large matrices.