The current MA standard for city/town names on street segments is to use one of the 351 official names as primary for streets. While this works well in most cases, there are situations (notably Boston’s neighborhoods and the Cape’s villages) where strict adherence to this rule can cause routing problems, and we have deviated from the rule in the past (particularly on the Cape).
To address this issue, the following change is proposed for street segments only (does not apply to Places):
Use official name as primary, unless:
There is a well-established village/neighborhood/CDP name, with defined boundaries, that is more commonly used (and has been cleared with SMs and listed in the MA wazeo as accepted). This can be used as the primary, with official name as alternate. [Consideration: will official name as alternate still create routing issues?]
Even if the boundaries are not well-defined, there is duplication in street name / address between villages (e.g. Harwichport/West Harwich) or a neighborhood vs. city (think Dorchester/Boston), leading to routing issues which could be solved by using village/neighborhood name as primary. Again, clearance by SM required. [But do you re-do an entire neighborhood just to fix one of two problem streets? Or keep more targeted?]
Any other oddball cases. Discussion, agreement, and publishing on the MA wazeo required.
Please reply with questions, comments, concerns, etc. Thanks.
Mudge42, thanks for bringing this topic forward. I like the idea that there will be a discussion among editors concerning the towns that fall into this category. Although those areas are not in my usual editing area I look forward to following the discussion and learning from it. Listing these townships that are exceptions in the Wazeopedia will enable editing to be consistent between editors.
We forsee that there will not necessarily be full agreement on what should or should not be a village, but a major reason for the requirement for discussion and SM approval is that it may be easier to define rules based on what we’re trying to prevent rather than what we’re trying to accomplish.
We don’t want situations where drivers disagree with a village name being used
We don’t want adverse consequences from unnecessary village name in the city layer of the app/editor and false alerts from editing scripts
We don’t want back/forth editing changes between editors who disagree on village usage.
We DO want an improved driving experience, particularly when there are ambiguities on street/city combinations (common in Boston area where formerly independent municipalities were annexed by Boston).
Excellent topic Mudge42. And it isn’t just Boston and The Islands area where we have this problem. i.e. Chelmsford / North Chelmsford where I decided to find all the segments with North Chelmsford and change them to Chelmsford to match the official listing of town / city names in Wazeopedia. There was a short discussion there about a former editor taking the time to name many of them as belonging to North Chelmsford and here I was going and changing them back to North Chelmsford.
With respect to your initial post,
did my renaming all those segments cause routing issues? I don’t know to be honest. I haven’t seen an increase in URs in that area, but that doesn’t mean that there haven’t been issues.
I seem to recall some discussion about this before somewhere, but could we use the USPS ZIP code definitions as a guideline?
In particular in Boston this eliminates many/most of the neighborhoods (South End, Back Bay, North End, to name a few, are not recognized USPS names), but captures the ones that I see the most (JP, Charlestown). In my experience, the USPS addressing conventions are how people list addresses for web sites, etc, so they’re the most likely to be in people’s minds. They also have well-defined boundaries.
I don’t know if the USPS data is appropriately licensed for us to use, or if there’s a good way to overlay it (I also seem to recall someone making a perhaps-unofficial overlay at some point?), but it’s an idea if it’s feasible…
As an example that just came up in Discord, there’s a UR (https://www.waze.com/editor?env=usa&lon=-71.77965&lat=42.19802&zoom=6&mapUpdateRequest=7863123) referring to 02130 and wondering why it doesn’t come up with Jamaica Plain - if you look up 02130 on the USPS site, Jamaica Plain is the primary name, with Boston as an “also accepted name”. If you look up my ZIP code in the South End, 02118, Boston is the primary name, with Roxbury as an “also accepted name”. My suggestion would be that the USPS primary name is what we would use - so nothing in Roxbury would change (which is good, because labeling things in the South End as Roxbury will DEFINITELY get us angry URs :)), but JP, Charlestown, and perhaps others (I haven’t exhaustively checked all of Boston’s ZIP codes :)) would get their more common usage.
If you read down into the weeds, even zip codes are not a pure representation, but they are as pure as we have.
Proposal: Pick a street segment in Jamaica Plains and change the primary city to Jamaica Plains. (Any city with a similar issue would do, but why not start here?)
Base the determination of primary and alternate WME names on the USPS designation of Default and Other.
After the tiles update, perform two searches for an address on that street: one using the Primary name and one using the Secondary. They should both return to the same location.
I know this works because I live in the same environment, but believe this experiment is necessary so that NER editors gain confidence that this works as a solution.
I would agree Zip Codes are a simple boundary as they often reflect neighborhoods names and as twilde42 indicated it is the way people list their address.
A quick comparison of different official resources on Neighborhoods for Boston reveals
I suggest that allowing the usage of these above BOLDED Neighborhoods in Boston (but by no means am suggesting a rapid change to the existing road network, but allow them if issues arise). Looking at the City layer in WME I already see that Charlestown & Hyde Park is established despite not being a recognized city, so do we go and process these two to be removed from existence or accept that they exist and acknowledge the other neighborhoods of Boston so as issues come up we can more efficiently address?
I agree that Banished test represents a good first step, I have done similar tests with alt road names for addresses (there was a Discord Presentation on alt names road HN). You might have to wait a couple of tile updates to ensure a new City Name takes hold (and you need a R4 to create the City), or test this out in one of the established Neighborhoods Charlestown/Hyde Park.
Thanks to all for the comments and suggestions. I have updated Boylston St in JP to have “Jamaica Plain” as the primary city, and “Boston” as the alternate (did not require adding Jamaica Plain as a new city name). After the tile update, we can test.
An option other than USPS zip codes (albeit slightly more complicated) which will allow generalizing beyond Boston - There is a Massachusetts official listing of streets called the MSAG (Master Street Address Guide) which was developed to provide uniqueness of Community and street names for 911 purposes.
“In most cases an MSAG community matches the boundaries of the 351 official municipalities in the commonwealth. Some cities and towns are divided into more than one MSAG community.”
“Therefore there are 383 MSAG community polygons in the state. Ten municipalities are further subdivided into more than one MSAG community. They are Barnstable, Boston, Deerfield, Dennis, Harvard, Nantucket, Northampton, Palmer, Shirley and Springfield.”
Caveats:
“The spelling of MSAG communities does not always agree with the official spelling of the town name. For example the town of North Attleborough is associated with the MSAG community of North Attleboro. Where towns are subdivided into multiple MSAG communities they are not a representation of villages or ZIP Code areas. Although MSAG community values are often the same as village names they do not represent a comprehensive listing of the villages for that town nor are they the limits of villages where the names agree. The MSAG was based on keeping the combination of street names unique and only where there was ambiguity was a MSAG community created.”
(2) For Boston I believe the USPS Zip and MSAG are pretty similar except MSAG removing Roxbury Crossing (lumped into Roxbury), Dorchester Center (lumped into Dorchester), and Allston (lumped into Brighton).
(3) I’m not sure if the proposal at the moment is to bulk rename neighborhoods. Another option is to address street name discrepencies individually as we come across them - either by only changing the city name primary/alt for those where we’ve received reports/complaints (and the corresponding duplicate), or by appending the village/neighborhood name to the street name as in the Hyannis/Osterville example towards the bottom of the page here.
If he cant do the overlay it is very easy to create a link to the specific area in the map from WME Lat/Lon, which the WME GIS script should be able to do
I believe both an overlay script and a WME GIS would be possible. I’ve also created overlay scripts but not nearly as elegantly as MOMs…
Chestnut Hill is a funny entity, being a village made of up three different Municipalities. I suspect it’s not in the MSAG since there there were not duplicated street names which needed differentiation; MSAG only used “neighborhoods” of “villages” when the municipality+street name was not unique.
Does this mean that Jamaica Plain has been accepted as an officially recognized City for Waze and should be added to the State wiki? What is the stance on the currently existing cities of Charlestown & Hyde Park?
We’ve had 3 tile updates since the test has started (maybe 4 depending on when the change was made), I tested 162 Boylston St Boston vs 162 Boylston St JP. Both HN exist on Boylston St in the different respective areas.
Most HN in Boylston St (Boston) have not been updated in the range that overlaps with JP which means this would not be perfect for testing as we would not be dealing with variables like waze interpolation and external sources.
Since we see that waze can distinguish two different roads based on city, what other criteria are we looking for in this “experiment”? What is the criteria to feel comfortable with this test that waze is appropriately routing and that JP and others can be an established City Name?
Yet the results in the app are not what I was expecting to see - 162 (and 194, which was in the UR that twilde mentioned) don’t come up as immediate suggestions - if you explicitly enter “xxx Boylston St Jamaica Plain”, the search suggestion is “xxx Boyston St, Jamaica Plain, Boston” - doesn’t sound like a Waze result. I’d like to wait and see.
Looking back I was maybe not clear, but the question was about how to handle the occurrence of TWO 162 Boylston St in Boston MA? Reviewing Massachusetts Interactive Property Map for JP & Boston exist, and now exist in Waze for House Numbers.
With 194 we are ultimately waiting on the indexing to be complete. Are we waiting for confirmation of indexing of the house number 194 or should we look more at the issue of duplicate addresses?
I am looking for a result in the app that will be clear for the end user. Right now I’m not seeing that for either 162 or 194 Boylston St, as the JP address is not readily appearing as a choice in the app search. And the result that you eventually get to doesn’t appear to be coming from Waze’s HN.
I would like to see some evidence that the trouble we will be going through to re-label segments is going to give us a better result than we have today. Surely we can wait a little longer for that evidence.