NaPTAN
Creation of National Public Transport Gazetteer (NPTG)
Overview
The National Public Transport Gazetteer (NPTG) is a common resource for use by all the traveline regions. Initially its primary use is to enable the call centre agents to determine which traveline region an origin and/or destination lies in. When JourneyWeb is used to link the regional journey planners together it will be used to automatically choose the appropriate Journey Planner for the distant parts of the journey. NPTG will also form an important element of the National Public Transport Access Nodes (NaPTAN) Database which will be used in traveline and the bus registration process run by Traffic Area Network (TAN).
The source material for NPTG is taken from the Index of Place Names compiled by ONS (and its Scottish equivalent). Whilst not completely ideal for traveline purposes these indexes are the most comprehensive list of settlements (which can range from hamlets to cities) covering England, Wales and Scotland.
This note describes the process by which the NPTG is to be compiled, using the ONS Gazetteer (and its Scottish equivalent) as the primary source of raw data. The mechanics of this work involve local database managers editing their own area's entries in accordance with a set of rules which are set out in this note. The work will be done over the internet - the source data is held on a computer managed by Thales, and you will be creating the PT Gazetteer alongside that source data by removing, correcting and adding entries as necessary - and then by adding PT-specific information to those entries.
When this process is completed, we will have created a National set of data that will be available for use in all of the traveline systems. It will be used to allow calls to be transferred to the relevant distant call centre where necessary - and it will also be used when local systems adopt JourneyWeb to obtain information from distant regions. It is already in use within the national traveline web portal.
It is important to remember that NPTG will be used by people all over the country, many of whom will not be familiar with places outside their region, and therefore every entry in NPTG has to be unambiguous. The aim of the editing process is to ensure that NPTG contains a single entry for each settlement name in the country, and ensure that the name, in conjunction with the county or unitary authority name, is unambiguous within the country.
An "alternative name" facility has been added to the editing process so that English and Welsh versions of place names can be tied together in NPTG. This facility can also be used for alternative spellings of place names and for places with more than one name such as Kingston-upon-Hull which is commonly known as Hull.
You are reminded that the use of the source material is governed by a legal agreement held by Intrasol Ltd (on behalf of traveline). Please do not breach the terms of this agreement by using the data for any other purposes without Intrasol's written agreement.
Guidelines for Settlements to be Included in NPTG
NPTG should include the names of all settlements, whether or not they are served by public transport. A settlement is defined as a distinct geographic area with a name that distinguishes it from adjacent areas. The names should be of importance at the national level, and should generally avoid small housing estate names (Remember that the purpose of the National Gazetteer is to allocate the journey to the region, although you also need to be able to allocate every individual bus stop in your area to an appropriate lowest-level gazetteer locality within the NaPTAN database).
Call centre agents and the general public are more likely to be familiar with county and unitary authority names than with second tier authority names (particularly those that are abstract) so every entry must be made unique at the county/unitary authority level. Where there have to be duplicate names within a county each must be qualified with a "(near This Place)" so that the agent/public can choose the right one.
The source material had multiple entries for large settlements that spanned several administrative areas. eg there were entries for "Leicester" in "City of Leicester" and "County of Leicestershire". The use of such multiple entries in NPTG will confuse call centre agents and the public. Such entries should be reduced to one (the one covering the centre of the settlement).
Some original entries contained more than one place name eg "Raynes Park/Motspur Park". These should be edited into two separate entries. (See the example edits below for other changes that should be made)
Hierarchy of Names
There can be a hierarchy of names (see editing activity 2 below) eg "Southampton" can include "Portswood" and "Shirley" as lower level names. The phrase "Town Centre" (and similar) should not be part of any highest level name in an area, although it can be used as a lower-order locality to represent the relevant part of a town. If, when creating the entries for a particular area you think you need more than 3 levels in the hierarchy , you are almost certainly trying to use too fine a settlement level. Also be careful not to create circular hierarchies. Hierarchies are to show subdivisions and are not intended to be used to show nearby, but distinct settlements.
The separate hierarchy editing system relates parent and child entries.
An example hierarchy
Alternate names
The National Public Transport Gazetteer has been extended to add the concept of alternate names for places known by several names eg English and Welsh names for the same place or a place such as Kingston-upon-Hull commonly known as Hull.
Each name and alternate name are entered as separate entries in the National Public Transport Gazetteer and then the alternate name editing system is used to tie the two entries together.
The hierarchy of place names will only be able to descend through a single name to simplify the construction of hierarchy trees but in all other respects alternate names will have the same status as the first name.
To ensure the integrity of hierarchy trees a locality can have the status of "Parent" or "Alternate", but not both.
An example of hierarchy and alternate names
The Alternate name editing system is very similar to the hierarchy editing system, but with a much smaller matching radius.
Places of National Interest
Places of national interest should only be included where the name does not make clear which settlement it is in. eg "Legoland" could be included as there is no obvious settlement name but "Chessington World of Adventure" should be excluded as there will be an entry for "Chessington". (Remember that the final part of the drill down process in the call centre will be using the local traveline gazetteer which will include all places of significance within the area).
Reports of ambiguous names found either within an administrative area or found close to an administrative area boundary can be provided to help the editors tidy up their part of the gazetteer. If required, please contact the Help Desk.
The Process
The gazetteer editing system will be accessible through the web page www.dft.gov.uk/public-transportdatamanagement/ng.htm. Access to the "live" system will be restricted to individual usernames and passwords which will limit access to a relevant geographical (administrative area) portion of the Gazetteer. Usernames and passwords will be issued on request to the Help Desk (those already in use remain valid).
There are three separate steps to follow.
Step One allows the local editor to refine the entries to be included in the NPTG and to indicate which of them have subordinate localities within the dataset (eg: a "town" which contains separately identified "suburbs").
Step Two allows the hierarchical relationships to be defined. A similar process will be developed for alternate names.
Step Three, using a graphical interface, defines links between localities and their stepping-off points for rail
Step One
The ONS entries will be listed alphabetically by Locality name, but viewed in blocks of 20 entries to minimise download times. The first column shows the District and may be required to assist in locating lesser-known, or duplicated, localities in large database areas. The third column, headed Typet is used to identify higher-order place names that have subordinate localities, or are alternate names for other localities within the database. The fourth column Code gives the ONS classification to each entry, as explained below. The fifth column, headed Active?, indicates whether the entry is active (= included in the NPTG) (4) or passive (= excluded from the NPTG) (6). The sixth column provides a Grid Reference which, if clicked on, will provide an on-screen map showing the position of that 1km grid square on an Ordnance Survey map base.
For each existing entry, there are a number of actions you can take —
To DELETE an entry, click on the tick in the 'Active?' column. It will change to a cross and the entry will be excluded from the NPTG (it remains in the original data-set and the deletion can be reversed if necessary)
To RENAME an entry, click on "Edit" in the 'Locality' column. A new window will open allowing you to change the spelling, add further words etc
To ADD a new entry, such as a missing suburb name or tourist venue, choose (as a "proxy") an existing entry geographically closest to the required entry (and in the same District as the required entry) and click on "New" in the 'Locality' column. A new window will open allowing you to overwrite the locality name of the proxy with an entirely different one. The Grid Reference should also be edited, as necessary. The entry which you used as the "proxy" remains unchanged by this process.
To nominate a locality as a PARENT (higher-order locality within a hierarchy), ALTERNATE (Alternate name for another entry in the locality table) or NEITHER change the radio button in the Type Column.
The head of the screen has buttons providing access to the main menu, to a sub-set of the list of Localities consisting of only those declared Active (View Included) and to a sub-set of the list of Localities consisting of only those declared Passive (View Excluded). Access to other ranges of the (View All) data is also available here. The titles of these ranges will automatically reset themselves as edits are made.
The Code column defines each source entry as:
- U (Urban Area)
- US (Urban sub-area)
- Pa (Parish – not Wales)
- Co (Community – Wales only)
- Lo (other Locality)
- DWD (Scottish District Ward – Scotland only)
- LOC (Scottish Locality – Scotland only)
- PAR (Scottish Parish – Scotland only)
- RED (Scottish Registered Electoral District – Scotland only)
- ISL (Island)
A further category
- Add (Added Locality)
…is also used in the editing process for entries which were not in the original source material.
The source data was devised to fulfil very different purposes to that which we are now applying it. Accordingly, and based on current experience, these codes provide the best method of sifting out unnecessary entries as follows:
| Review each entry coded U or US in "Code" column | |
|---|---|
| Entry is a single locality NOT already represented by another entry | No action |
| Entry is a single locality already represented by another entry coded Pa, Co or Lo | DELETE |
| Entry links two or more separate locality names of which at least one is NOT already represented by another entry | ADD and, probably, DELETE |
| Entry is two or more separate localities each of which is already represented by another entry coded Pa, Co or Lo | DELETE |
| Despite its allocation to a relevant district, locality is actually outside the administrative boundary | DELETE |
| Review each entry coded Pa or Co in "Code" column | |
| Entry is an exact match for a settlement name, and that settlement does NOT appear as a separate entry coded Lo | No action |
| Entry is an exact match for a settlement name, and that settlement already appears as a separate entry coded Lo | DELETE |
| Entry links two or more separate localities of which at least one is NOT already represented by another entry | ADD and, probably, DELETE |
| Entry links two or more separate localities each of which is already represented by another entry coded Lo | DELETE |
| Parish or Community name does not include any recognised settlement name | DELETE |
| Review each entry with a parenthesis) after locality name | |
| Two adjacent entries are identical, though in different districts. Grid references will also match | DELETE one entry |
| An entry followed by ) is NOT the same locality name as adjacent entries. Locality also appears in the gazetteer for an adjacent authority. | No action, but liaise with neighbouring authority |
| Then, review each remaining entry to identify other necessary amendments, looking particularly for the following categories | |
| Where a locality name consists of two or more words, ensure that the "forward" entry (Great Horwood) exists and that the "reverse" entry (Horwood, Great) does not | ADD and/or DELETE as appropriate |
| Localities including the word "and". Consider the need to split to separate localities | ADD and/or DELETE as appropriate |
| Adjacent entries have the same locality name but are attributed to different districts. Grid references confirm that these are two separate localities | RENAME by adding "(near <wherever>)" after name |
| Adjacent entries have the same locality name but are attributed to different districts. Grid references confirm that this is a single locality close to, or straddling, the boundary | DELETE one entry |
| Check for irregular spelling of, or (de)merging of words in, place name | RENAME to local spelling |
| Check for insignificantly small localities. Occasional occurrences found of localities named after roads or single buildings. | DELETE |
| Some localities listed will be recognised as having no public transport services | DO NOT DELETE. Instead instigate "walk" link within database or otherwise link to nearest locality(s) served by public transport |
| Then, consider the requirement for additional entries | |
| Review provision of area names, particularly within urban areas to establish omission of well-used locality names | ADD as appropriate |
| Is the place known by other names which appear in a different part of an alphabetical list. In Wales, consider inclusion of Welsh language form of place names | ADD as appropriate. English and Welsh place names will have to be tied together using the alternate name process (to be developed) |
| Consider tourist venues or other "places of interest" of national significance. Only include places that give added functionality. (eg include "Blenheim House" which is in Woodstock but exclude "Harlech Castle" because "Harlech" will already exist as a locality). The number of additions should be small; they will be monitored and may be reduced by negotiation. | ADD as appropriate |
The final part of Step One is to identify those localities that are likely to be the higher-level localities within a hierarchy or are alternate names. Use the radio buttons to change the status. This change will be saved automatically.
Changes in the edit window are saved when the "Update" button is pressed.
To check on finished selections, use the "View Included" or "View Excluded" buttons to view complete lists of Active/Included localities or Passive/Excluded localities. Return to Main Menu to commence Step Two.
Step Two A (Hierarchy)
This is the process by which the dependent localities (or 'children') are determined for each nominated 'Parent'. A list of localities will be presented for each 'Parent', listed in ascending distance from the parent locality. A simple toggle switch allows editors to include as many of the localities offered as is appropriate. (The default state for each is "excluded"). A trap is included to prevent simple (parent-child-parent) circular references being included.
The changes made during Step Two A are saved when "Update database" button is pressed.
Step Two B (Alternate Names)
This is the process by which the alternate localities are tied to their other name. A list of localities will be presented for each 'Alternate', listed in ascending distance from the Alternate locality. A radio button switch allows editors to select one name as the other name. The other name must not have been marked in the localities editing as an alias name.
The changes made during Step Two B are saved when "Update database" button is pressed.
Return to Main Menu to commence Step Three.
Step Three – Requires version 4, 5 or 6 of Netscape or Internet Explorer
The purpose of this step is to associate each locality with a range of railway stations and coach stops that form its national stepping-off points (points at which transfers may be made to the national rail or coach networks for long distance journeys). To minimise the time taken by this process, an "area of influence" is drawn for each stepping-off point and that point is allocated to all localities within the defined area. Choose Step Three on the Main Menu and a graphical representation of your area is presented on screen. A north point clarifies the true orientation. Press F11 to maximise the size of the image on your screen. A zoom option is available and, at enlarged scales, the use of scroll bars may be required.
If you see an image with a small cluster of points in the centre of the screen and a number of outlying points there are problems with coordinates of the outlying points. Try and identify the outliers by putting the cursor over the point which will show the name. Correct the coordinates using step one before proceeding with step three.
Those localities selected in Step One as "Active" are shown as black crosses. As you pass your cursor over the plot, the names of the localities appear.
To start the process, press "Select Rail Station" and all railway stations within the area of the plot are shown as blue dots. Alternatively, pressing "Select Coach Stop" will display all coach stops within the area as green dots. In both cases, by passing the cursor over a dot the name appears in blue (rail station) or green (coach stop). In either case, click on one of the relevant stepping-off points. The dot will turn red and its name will be more permanently displayed. Now press "Start Region" and click on one of the localities that looks to that station as a stepping-off point. Moving clockwise click on other localities and you will find that a boundary line is drawn with all the localities contained within it turning red. Do not worry if other localities turn red. As you complete the boundary by clicking back on the first locality again, only those localities within the boundary will remain red. If at any point you are happy with the localities marked in red, you can press End Region and this will complete the boundary for you – use this option with care, though! If, after completing the region, you find you have made a slight mistake and wish to add additional localities outside the boundary, or to exclude localities within the boundary, press "Start Edit" and click on those localities you wish to amend. When you are satisfied, press "End Edit". If you get it badly wrong, press "Cancel Region" and start that area again. Then press "Update Registry" to capture the data and to allocate that stepping-off point to all of the marked localities.
Repeat this process for each appropriate stepping-off point. If you click on a stepping-off point you have already dealt with, its "area of influence will be shown by the localities turning red. Once again you can press "Start Edit" and make further changes. Remember to "Update Registry" before moving on to another stepping-off point.
A procedure is being developed which will allow editors to check that all Gazetteer entries within their area have at least one nominated stepping-off point for both Rail and Coach modes, and to allow editors to check an individual locality's nominated stepping-off points.
Completion
Once you have completed Step Three, return to Main Menu and exit the application. Notify the Help Desk (Thales support) that you have completed your component of the National Gazetteer.
Updating
The editing system remains active at all times to allow continual editing after the initial work has been completed. The NPTG files that are required by end-users of the data are compiled at regular (monthly) intervals for distribution to traveline systems integrators and other authorised users.
APPENDIX 1The following table shows around 40 of the 691 entries in the Buckinghamshire data and provides examples in the repeated table of review actions on the following sheets.
| District | Locality | Type | Code | GridRef | |
|---|---|---|---|---|---|
| Aylesbury Vale | Addington Edit New | oParentoAlternateoNeither | Pa | a | 474228 |
| Chiltern | Amersham Edit New | oParentoAlternateoNeither | Pa | a | 496197 |
| Chiltern | Amersham Edit New | oParentoAlternateoNeither | US | a | 496197 |
| Chiltern | Amersham/Chesham Edit New | oParentoAlternateoNeither | U | a | 496199 |
| Aylesbury Vale | Biddlesden Park Edit New | oParentoAlternateoNeither | Lo | a | 463239 |
| Aylesbury Vale | Bierton Edit New | oParentoAlternateoNeither | Lo | a | 483215 |
| Aylesbury Vale | Bierton with Broughton Edit New | oParentoAlternateoNeither | Pa | a | 484215 |
| Aylesbury Vale | Blackgrove Edit New | oParentoAlternateoNeither | Lo | a | 476218 |
| Chiltern | Bois, Lower Edit New | oParentoAlternateoNeither | Lo | a | 496200 |
| Aylesbury Vale | Broughton Edit New | oParentoAlternateoNeither | Lo | a | 484213 |
| South Bucks | Burnham Edit New | oParentoAlternateoNeither | Lo | a | 493183 |
| South Bucks | Burnham Edit New | oParentoAlternateoNeither | Pa | a | 493183 |
| Wycombe | Chepping Wycombe Edit New | oParentoAlternateoNeither | Pa | a | 489191 |
| Chiltern | Chesham Edit New | oParentoAlternateoNeither | Pa | a | 496201 |
| Chiltern | Chesham(Amersham/Chesham) Edit New | oParentoAlternateoNeither | U | a | 496199 |
| South Bucks | Denham Edit New | oParentoAlternateoNeither | Pa | a | 503187 |
| Aylesbury Vale | Denham Edit New | oParentoAlternateoNeither | Lo | a | 475220 |
| Wycombe | Desborough Edit New | oParentoAlternateoNeither | Lo | a | 484193 |
| Wycombe | Desborough & Desboro' Castle Edit New | oParentoAlternateoNeither | Lo | a | 484193 |
| Wycombe | Desborough Castle Edit New | oParentoAlternateoNeither | Lo | a | 484193 |
| Aylesbury Vale | Dunsmore Edit New | oParentoAlternateoNeither | Lo | a | 486205 |
| Wycombe | Dunsmore Edit New | oParentoAlternateoNeither | Lo | a | 486205 |
| Aylesbury Vale | Eaton Bray Edit New | oParentoAlternateoNeither | U | a | 497219 |
| Wycombe | Great and Little Kimble Edit New | oParentoAlternateoNeither | Pa | a | 481207 |
| Wycombe | Great Kimble Edit New | oParentoAlternateoNeither | Lo | a | 482205 |
| Aylesbury Vale | Little Horwood Edit New | oParentoAlternateoNeither | Pa | a | 479230 |
| Chiltern | Little Kingshill Edit New | oParentoAlternateoNeither | Lo | a | 489199 |
| Chiltern | Lower Bois Edit New | oParentoAlternateoNeither | Lo | a | 496200 |
| Aylesbury Vale | Stone Edit New | oParentoAlternateoNeither | U | a | 479212 |
| Situation | Action(s) to take | Example in Buckinhamshire |
|---|---|---|
| Review each entry coded U or US in "Code" column | ||
| Entry is a single locality NOT already represented by another entry | No action | Retain Stone (U) because it is not mirrored by a similarly named Pa or Lo entry |
| Entry is a single locality already represented by another entry coded Pa, Co or Lo | DELETE | Delete Amersham (US) because of presence of Amersham (Pa) |
| Entry links two or more separate locality names of which at least one is NOT already represented by another entry | ADD and, probably, DELETE | No example found |
| Entry is two or more separate localities each of which is already represented by another entry coded Pa, Co or Lo | DELETE | Delete Amersham/Chesham (U) because of presence of both Amersham (Pa) and Chesham (Pa) |
| Despite its allocation to a relevant district, locality is actually outside the administrative boundary | DELETE | Delete Eaton Bray (U) because the settlement is actually in Bedfordshire |
| Review each entry coded Pa or Co in "Code" column | ||
| Entry is an exact match for a settlement name, and that settlement does NOT appear as a separate entry coded Lo | No action | Retain Addington (Pa) because it is not mirrored by a similarly named Lo entry |
| Entry is an exact match for a settlement name, and that settlement already appears as a separate entry coded Lo | DELETE | Delete Burnham (Pa) because of presence of Burnham (Lo) |
| Entry links two or more separate localities of which at least one is NOT already represented by another entry | ADD and, probably, DELETE | Consider Great and Little Kimble (Pa). Great Kimble (Lo) already exists but Little Kimble (Lo) does not. Use NEW to create Little Kimble (Add) from copy of Great Kimble (Lo), editing name and grid ref accordingly |
| Entry links two or more separate localities each of which is already represented by another entry coded Lo | DELETE | Delete Bierton with Broughton (Pa) because of presence of Bierton (Lo) and Broughton (Lo) |
| Parish or Community name does not include any recognised settlement name | DELETE | Delete Chepping Wycombe (Pa) as it is not the name of a locality |
| Review each entry with a parenthesis) after locality name | ||
| Two adjacent entries are identical, though in different districts. Grid references will also match | DELETE one entry | Dunsmore (Lo) appears twice, attributed to both Aylesbury Vale and Wycombe districts. Village straddles boundary. Delete one occurrence |
| An entry followed by ) is NOT the same locality name as adjacent entries. Locality also appears in the gazetteer for an adjacent authority. | No action, but liaise with neighbouring authority | Biddlesden Park (Lo) listed. Park extends across county boundary into Northamptonshire. Entrance/access will determine which county (or both) should retain the entry |
| Then, review each remaining entry to identify other necessary amendments, looking particularly for the following categories | ||
| Where a locality name consists of two or more words, ensure that the "forward" entry exists and that the "reverse" entry does not | ADD and/or DELETE as appropriate | The entry Lower Bois (Lo) does exist and so does Bois, Lower (Lo). Delete Bois, Lower (Lo) |
| Localities including the word "and". Consider the need to split to separate localities | ADD and/or DELETE as appropriate | Delete Desborough and Desborough Castle (Lo) because of separate entries for Desborough (Lo) and Desborough Castle (Lo) |
| Adjacent entries have the same locality name but are attributed to different districts. Grid references confirm that these are two separate localities | RENAME by adding "(near <wherever>)" after name | Denham (Lo) in Aylesbury Vale changed to Denham near Quainton and Denham (Lo) in South Bucks changed to Denham near Uxbridge |
| Adjacent entries have the same locality name but are attributed to different districts. Grid references confirm that this is a single locality close to, or straddling, the boundary | DELETE one entry | Dunsmore (Lo) appears twice, attributed to both Aylesbury Vale and Wycombe districts. Village straddles boundary. Delete one occurrence |
| Check for irregular spelling of, or (de)merging of words in, place name | RENAME to local spelling | Foscott (Lo) is actually spelt Foscote |
| Check for insignificantly small localities. Occasional occurrences found of localities named after roads or single buildings. | DELETE | Blackgrove (Lo) does not appear as a placename in Royal Mail's postal Address Book. It can only be traced on any published map as "Blackgrove Farm House". Therefore delete as insignificant. |
| Some localities listed will be recognised as having no public transport services | DO NOT DELETE. Instead instigate "walk" link within database or otherwise link to nearest locality(s) served by public transport | Green Hailey (Lo) is a valid locality, being a clutch of houses. It is not served by any public transport services and Princes Risborough will be identified as the closest place served. |
| Then, consider the requirement for additional localities or "Places of Interest" to be included within gazetteer | ||
| Review provision of area names, particularly within urban areas to establish omission of well-used locality names | ADD as appropriate | Haydon Hill is a well known part of Aylesbury but does not have an entry in ONS's IPN |
| Is the place known by other names which appear in a different part of an alphabetical list. In Wales, consider inclusion of Welsh language form of place names | ADD as appropriate | No such examples in Buckinghamshire! |
| Consider tourist venues or other "places of interest" of national significance. Only include places that give added functionality. (eg include "Blenheim House" which is in Woodstock but exclude "Harlech Castle" because "Harlech" will already exist as a locality). The number of additions will be monitored and may be reduced by negotiation. | ADD as appropriate | Stoke Mandeville Hospital is the home of the National Spinal Injuries Centre. It is in Aylesbury, not in Stoke Mandeville itself. |
Page last updated: 28.06.2005
