Legal description of Hudson's city boundaries

[Kevinstat]

Would the building on the northern corner (or perhaps "northwest corner in Hudson-speak) of 3rd Street and Robinson Street (southeast of where Strawberry Alley curves to hit Robinson Street) be in your (A)?  Or is it already in its own (planned by the Census Bureau) census block, separated from the bureau's planned (I) by the entrance from the end of 3rd Street to the cleared area on the north(-northeast) side of Strawberry Alley?  Your description of the Census Bureau's planned (I) made it seem like it would include that building.  It looks like it might be a residence (Google Maps lacks street views for a surprising amount of Hudson).  My guess is that building on the north side of Strawberry Alley is just a storage shack, with no residents.  If so, than the building on the corner of 3rd and Robinson is separate enough from your (A) that using that northern entrance as a block boundary if possible might be worthwhile.

[Kevinstat]

It is actually a quirk, that the houses on the north side of Robinson are a separate census block.

I'm afraid I don't understand what you're saying here.  The alleys in Hudson (and of course the streets) seem to usually be census block boundaries, as they are in this case.

There appear to have been past efforts to carve out census blocks in the area. For example, earlier maps show a cut from Rope Alley to State Street at about 4th Street, which would make a block of houses on the north side of State Street, and there are some other lines that only make sense if someone was trying to bring things back into the city.

I don't see such a block boundary looking at the 1990 (the Columbia county block map covering the less built-up areas of all but the southernmost part of Hudson; if you're interested, you could see the county block maps covering "inner city" (inset) and southernmost Hudson), 2000 and 2010 census block maps for Hudson.  You might be looking at earlier maps not available online, however.

[Kevinstat]

Looking at the 1990 census block map covering southernmost Hudson (where the boundary between Hudson and Greenport follows Ten Broeck Lane) and comparing that to the 2000 and 2010 census block maps, I can see that the Census Bureau has been willing to adjust its census tract boundaries in the past, and changing block group boundaries (I know both the census tract and block group boundaries have been mentioned on this thread, like in the map you made with your A, B and C that you colored by block group).  Adjusting the census tract boundary to follow the (at least rough) extension of Warren Street to the railroad tracks (or the city/county line in the Hudson River) as some might wish might be a different story, however.

[Kevinstat]

Have you downloaded a copy of QGIS?

Not yet.  I probably will tonight or some evening this coming week.  I've been using the free PC version of ArcGIS.  A lot of the shapefiles the census bureau provides may not be that useful for me, like school districts.  After Maine's penalized if you don't do it consolidation under Baldacci, some districts were formed with old districts far enough away from each other that they wouldn't lose their high schools.  I could get more accurate information about school ties between towns from information other than the Census Bureau.

The replies I've received so far from Maine's Redistricting Data Program liaison (Amanda Rector, also the State Economist), have been fairly pro forma.  Maybe that's not the right word, but you know, nothing like "I've looked into your suggestions about the block covering Boothby Street and...".  If I were to see a current Tiger line (not currently planned as a block boundary) that matched a division I wanted I'd definitely mention that, but overall I feel rather insignificant to the process right now.

Does the presence of an old Tiger line (like if you wanted that area between Rope Alley and State Street (north)west of roughly 4th Street to be in a separate block from the rest of it's visual block) make things any easier as far as adding new lines?  If not, then I'll probably just use the 2016 shapefiles (or really fall 2015 (V2), but in one page the Census Bureau calls it 2016).

Augusta has 5 census tracts, three west of the river, divided by Western Avenue and Bond Brook; and two east of the river divided by Cony Street and South Belfast Avenue.

Augusta has 17 block groups, and a census tract will typically have a couple in the built-up areas, and one or two in the outlying area.

Have you noticed the odd jog in the block group boundary in the southeast of the west side of Augusta between Sewell and State Streets along South Street?  I think at one point Sewell Street may have ended on South Street, which extended a bit further to the west, rather than continue into Hallowell.  I saw a map in an online news article about some "warmed cold case" homicide from (the homicide) 1979 IIRC (around 1980 at least).

Kennebec County Census Tract 010801, which is Manchester, is displayed as "108.01". The fraction indicates that this is a modified version of "108". Tract 108.02 is Farmingdale and West Gardiner. So at some time, all three towns were in tract 108. They keep the 108, to identify that they shared a common predecessor. If Farmingdale and West Gardiner were split, then 108.02 would be retired, and the new tracts would likely be 108.03 and 108.04.

Incidentally, the split of 108 was asymmetric, since Manchester is the smallest of the three. But it probably makes the most sense from a compactness sense, and direction from Augusta, with Manchester to the west, and the other two to the south.

The split of Tract 108 happened between the 1990 and 2000 censuses.  Maybe the other two in Kennebec as well, but I don't have time to check right now.

There have been two other splits in Kennebect County. Waterville has gone from two to three tracts: 241.01, 241.02, and 242; and Winslow has been split 230.01 and 230.02.

Is the northern part of the county around Waterville more dynamic?

Hmmm.  Waterville had 2+ State House districts after the 1990 census, but now has 2-, although census estimates projected out to 2020 (using absolute Arp. '10 to Jul. '15 gains multiplied by 10/4.25) has the city gaining relative to the state now and passing the 1.90/151 mark that would allow it to have two whole districts (when you have 151 districts to draw, +/- 5% is a good rule of thumb as you don't know what other districts' populations will be when you start drawing them.  Winslow had 1+ districts in the 1980s apportionment, then 1, then needed a bit of Benton, and now more of Benton (the two-phase gains in Benton have also been the only changes (losses in this case) of the China-Albion-Unity Township-(majority of) Benton district.

[Kevinstat]

You may want to download QGIS:

QGIS download

In the "Latest release (eg. for New Users)" section alone there are four options, 32 bit or 64 bit and for each of those a .exe and a .exe.md5sum .  What should I download?  I have an HP Split 13x2 Detachable PC with Windows 10 (it came with Windows 8.1 but I took advantage of the free upgrade, although in some ways I liked the 8.1 better).

I copied our discussion regarding Maine to my Augusta (mostly), ME census block/ward/house district boundary issues thread, btw.

[Kevinstat]

Looking at the 1990 census block map covering southernmost Hudson (where the boundary between Hudson and Greenport follows Ten Broeck Lane) and comparing that to the 2000 and 2010 census block maps, I can see that the Census Bureau has been willing to adjust its census tract boundaries in the past, and changing block group boundaries (I know both the census tract and block group boundaries have been mentioned on this thread, like in the map you made with your A, B and C that you colored by block group).  Adjusting the census tract boundary to follow the (at least rough) extension of Warren Street to the railroad tracks (or the city/county line in the Hudson River) as some might wish might be a different story, however.

I think they consider that a location correction. ...

Census tracts are intended to be used for statistical purposes. The first were delineated in New York City in 1910. Local groups could set the boundaries, and the census would enumerate the data based on those boundaries. So you  could get the same level of information for Tract 123 in Brooklyn, as you could for Manchester, Maine. Otherwise you would have the same level of detail for the entire borough of Brooklyn and Manchester, Maine.

Data for census tracts was previously derived from the long form, and now the ACS, and is based on a sample, and the intent is that they can be compared over time.

So I take it my idea of the extension of Warren Street to the Hudson River (what Torie is proposing for a ward boundary in the other Hudson thread) as a census tract boundary (as opposed to a set of block boundaries) would be a no go.  Correct?

Do your A and B in the map below have any residences in them?  If not, then your idea of them moving between blocks (rather than having to create new ones) might be acceptable to the Census Bureau.  If either of them do, then you might have to go with your alternative of making that area its own census block in CT 13, BG 1.

Color codes are by block group.

Green is CT 12, BG 1;
Purple is CT 12, BG 2;
Pink is CT 13, BG 1;

My proposal is to eliminate Front Street north of Dock Street (non-existent); extend the delineation of Dock Street to the west to reach the RR tracks; and eliminate all of the imaginary lines except from the RR bridge over the estuary to the middle of the Hudson.

Area A would be added to 12-2000 (north of Dock Street to stream), between 2nd Street and RR tracks.
Areas B and C would be added to 12-2004 (North Bay estuary). Area C is moved to use west track of RR as the block line.

If the Census Bureau insists in maintaining the census tract boundaries, then the imaginary lines from Front Street north and west to across estuary would have to be maintained.

Area A would be a new census block in CT 13, BG 1.
Area B would be a new census block in CT 13, BG 1.
Area C would be moved from 12-1007 to 12-1004.

[Kevinstat]

You may want to download QGIS:

QGIS download

In the "Latest release (eg. for New Users)" section alone there are four options, 32 bit or 64 bit and for each of those a .exe and a .exe.md5sum .  What should I download?  I have an HP Split 13x2 Detachable PC with Windows 10 (it came with Windows 8.1 but I took advantage of the free upgrade, although in some ways I liked the 8.1 better).
...

I'd go with the 64-bit.  I think the md5 version has a checksum that can be used to check that you have an authentic version.
...

I clicked to download the md5 "version" of the QGIS Standalone Installer Version 2.14 ("for new users") 64-bit, which download was almost instantaneous, but I didn't have a program that seemed designed to download the resulting MD5SUM File.  So I clicked to download the regular QGIS Standalone Installer Version 2.14 64-bit, and its taking a few minutes just to download the installer.  That may be normal, or my computer may have some issues.

[Kevinstat]

You may want to download QGIS:

QGIS download

In the "Latest release (eg. for New Users)" section alone there are four options, 32 bit or 64 bit and for each of those a .exe and a .exe.md5sum .  What should I download?  I have an HP Split 13x2 Detachable PC with Windows 10 (it came with Windows 8.1 but I took advantage of the free upgrade, although in some ways I liked the 8.1 better).
...

I'd go with the 64-bit.  I think the md5 version has a checksum that can be used to check that you have an authentic version.
...

I clicked to download the md5 "version" of the QGIS Standalone Installer Version 2.14 ("for new users") 64-bit, which download was almost instantaneous, but I didn't have a program that seemed designed to download the resulting MD5SUM File.  So I clicked to download the regular QGIS Standalone Installer Version 2.14 64-bit, and its taking a few minutes just to download the installer.  That may be normal, or my computer may have some issues.

The download of the installer finished, but I deleted it after I agreed to run it, got one security prompt, said to run anyway, and then got another.  Why is the publisher listed as unknown?

[Kevinstat]

The census tract is also the basis to award certain federal funds based on the statistics for the tract. In my area there were a great number of census tracts that were split between 2000 and 2010, even in old established neighborhoods with little or no growth. From a local standpoint those splits could aid the application for funds.

When the Census made tract splits they got rid of the old number and created new tract numbers. That made the 2010 tracts more useful locally and avoid any confusion with past data sets. For example in 2000 I lived in 8415.02. There was minimal new construction in the tract during the decade, but it was split into 8415.03 and 8415.04. The old number went out of service for 2010 and beyond.

If Hudson wanted a change, the Census could also assign new numbers to the revised tracts as they did in my area in the last decade.

That appears to have been an ordinary split. 8415.02 had 6918 persons in 2000. 8415.03 and 8415.04 had 3923 and 3294 in 2010 (at total of 7217, or 4.3% growth).

Census tracts have 1200 to 8000 residents with 4000 being the ideal. Had there not been a split the tract would have been 180% of the ideal. With the split, the smaller tract is 82% of the ideal and the other is 98% of the ideal.

The Census Bureau does not define census tracts for the purposes of garnering federal grants. Congress is free to ignore their advice. The Census Bureau will do custom ACS tabulations.

In any case, it makes no census to run a census tract boundary through a complex of rent-subsidized apartments. For statistical purposes, you would prefer less variation within census tracts, and more variation between census tracts.

I thought I'd do this before you get one of those cases of multi-layer quotes where the quotes are all attributed to the wrong person (I deleted the inner-most quote-citation for jimrtex in the last post that caused that error).

Color codes are by block group.

Green is CT 12, BG 1;
Purple is CT 12, BG 2;
Pink is CT 13, BG 1;

From looking at your various posts on this topic, I gather your position shifting the census tract boundary in Hudson is as follows (and I'll throw in my take on your position on whether a proposed area should at least be divided by a block line for good measure):

Your C (doesn't involve a census tract change): you definitely support moving this area from 12-1007 to 12-1004.

Your B: you support moving this area from CT 13 to CT 12 (from 13-1000 to 12-1004), although perhaps not as emphatically as your support for the move of C.  Even if the census tract line is kept, you at least marginally support B being removed from 13-1000 (becoming its own block in this case).

Your A: you definitely support this area being removed from 13-1002.  You marginally prefer it remaining in CT 13 (becoming its own block in this case).  If it is to be moved to CT 12, you definitely support it being combined with (added to) 12-2000, and that's the only reason your support for its remaining in CT 13 is marginal (if I'm correct in that assumption) rather than strong.

Remaining area of CT 13 north of the extension of Warren Street: You emphatically oppose the idea of this area moving to CT 12.  You support this extension becoming a block line from Front Street to the more westerly of the two RR tracks.  Regarding that extension going beyond that, to the shoreline or beyond to the municipal/county line in the Hudson River, you're meh at best (perhaps less seeing the point of going beyond the shoreline).  You also may doubt the Census Bureau would do that, and you definitely doubt the Census Bureau would make any of the extension of Warren Street a revised Census Tract line.  I'm not sure if Muon2 realized that your strongest opposition was against this shift, rather than the shift of A and B (and indeed, you support the shift of B) above.