usaky3: Kentucky Highways 1000-1499

[mapcat]

I might do some GMSVing around until I get a good feel for what's more likely to match what's in the field.
In my very limited experience at the south end of KY US127BusAlb, KYTC names were more accurate.

Whenever there's a discrepancy I check GMSV. Trouble is that much of the GMSV in the eastern part of the state is very old and street signs are illegible.

[yakra]

1020 has very spotty signage throughout urban Louisville, but KYTC's route is close to what we have. Also, KYTC shows part of it along a one-way pair, so waypoints need to be re-centered.

I'd go with KYTC... Is there any signage you can find that outright contradicts this?Or is signage just... poor?

KY1059: extend to KY32 or not?

I'll let you be the judge of that. Anyone else got an opinion?
If this were Maine, I'd include the extension. Probably. But this isn't Maine, is it?

[mapcat]

1020 has very spotty signage throughout urban Louisville, but KYTC's route is close to what we have. Also, KYTC shows part of it along a one-way pair, so waypoints need to be re-centered.

I'd go with KYTC... Is there any signage you can find that outright contradicts this?Or is signage just... poor?

Louisville map (see sheet 3)

Signage exists at the intersection with US60A but is (almost) entirely missing north of it. Between there and I-264 there isn't any signage but it follows one roadway without turning. South of I-264, it turns several times through a residential area but the first signed turn I found is at the intersection of Southside Dr and National Turnpike, even though it has a rather convoluted routing along residential streets in between. From that point southward, it seems to be signed well.

I examined some more logical routes for it to take through the city but didn't find signs anywhere else nearby, either.

Since it's signed at US60A, I think we should at least keep it from that point south. North of there, its official route at least follows the same streets (2nd & 3rd) to its unmarked end, and according to some of my old official maps it has followed them since at least the early '90s (older maps didn't have as much detail). OTOH, the US routes and KY61 and KY864 are comparatively well-signed downtown. Right now I'm inclined to reroute it along the KYTC routing all the way to its official end, since I can't find any signage that conflicts with that, although deleting it north of US150 (the first signed route it encounters north of US60A) would make about as much sense.

[mapcat]

Here are the notes for the 11XX set:


Deleted or unknown:
1111
1143
1188
1190
1197


Unsigned but still KYTC-maintained:
1174


Incorrect endpoint locations:
1106
1145
1185


Incorrect endpoint/waypoint names:
1101
1102
1103
1110
1116
1117
1120
1122
1124
1132
1133
1140
1147
1148
1150
1151
1152
1153
1156
1165
1167
1169
1171
1175
1179
1181
1186
1187
1192
1198


Missing or extraneous waypoints (false intersections that can be eliminated without affecting shape):
1107 (missing KY1145)
1151
1153


The rest appear ok. Corrections will be made and files submitted over the weekend.


I have not begun checking for datacheck errors because I am unsure how to mark false positives.

[yakra]

I have not begun checking for datacheck errors because I am unsure how to mark false positives.

The code in the "FP Entry to Submit" column gets added as a new line to datacheckfps.csv in the root directory of the HighwayData repo. (The URL will be different when pointing to whatever branch you're editing.)

[Jim]

I have not begun checking for datacheck errors because I am unsure how to mark false positives.

The code in the "FP Entry to Submit" column gets added as a new line to datacheckfps.csv in the root directory of the HighwayData repo. (The URL will be different when pointing to whatever branch you're editing.)

A note on this: as long as we keep it more or less alphabetized different people's edits there will not conflict with each other, so it's safe to submit pull requests with commits to datacheckfps.csv.  Of course you can also put them in the text description of the pull or email them and I can add them.  Just pointing out that this situation is different from updates.csv, where we'd all be adding in the same place, which an automatic merge would not be able to handle (nor should it be able to) so we'd end up with conflicts to resolve when merging in those pull requests.

[mapcat]

The code in the "FP Entry to Submit" column gets added as a new line to datacheckfps.csv in the root directory of the HighwayData repo. (The URL will be different when pointing to whatever branch you're editing.)

Um...
When I look at the file the columns are:

Route;Waypoint1;Waypoint2;Waypoint3;Error;Info

[mapcat]

Notes for the 12XX routes, to be fixed over the weekend:



Deleted or unknown:


1264


Unsigned but still KYTC-maintained:


1225 (signed as "To 80"


Incorrect endpoint locations:


1201
1202
1205
1206
1227
1240
1242
1246
1258
1271
1273
1275
1281
1283 (end marked at 45; KYTC says it ends 1 block east)


Incorrect endpoint/waypoint names:


1203
1214
1220
1223
1230
1232
1234
1238
1241
1244
1245
1247 (town names for KY1247 aren't going to work since there are more intersections with US27 than there are towns)
1249
1250
1260
1267
1272
1273
1274
1277
1292
1295
1298


Missing or extraneous waypoints (false intersections that can be eliminated without affecting shape):


1268


Incorrect shaping points


1221

[mapcat]

Notes for the 13XXs, to be fixed this weekend:

Deleted or unknown:

1334
1341
1358
1370
1378
1386

Unsigned by KYTC-maintained

1381

Incorrect endpoint locations (most are small adjustments)

1308
1313
1318
1335
1345
1346
1347
1361
1371
1384
1388
1391 (both ends)
1392

Incorrect endpoint/waypoint names:

1303
1305
1311
1316
1326
1327
1328
1329
1331
1332 (also needs waypoint for US231)
1336
1345
1348
1353
1355
1356
1359
1363
1366
1367
1373
1379
1380
1384
1385
1389
1393
1396
1399

Reverse order to match mileposts

1301
1355
1399

Missing or extraneous waypoints (false intersections that can be eliminated without affecting shape):

1336
1340

Other

1309 (move shaping point to nearby BMarRd)
1376 (move FarRd)

KY1303 - KY1829_W on Industrial Road according to tile mapping.
KY1312 - Cum/Met not on county boundary from open source data (MQ Open/Mapnik)
KY1313 - Rus/Ada not on county boundary from open source data (MQ Open/Mapnik)
KY1316 - MosCreRd -> KY2352
KY1377 - KY80 -> KY80/1371 (and a similar move on KY1371)
KY1384 - US23/119_N - doesn't meet US119, does meet US460

1303: noted above
1312 & 1313: the endpoint matches the endpoint (county line) indicated in KYTC aerials, so this is unnecessary
1316: KY2352 is unsigned, so this is unnecessary
1371 & 1377: these routes don't intersect KY80 in the same location, as indicated in the KY80 file, so this is unnecessary
1384: noted above; the KY1460 end is also at the wrong location

[mapcat]

Notes for the 14XXs, which I hope to fix this weekend.


Deleted or unknown:


1423
1467
1474
1478
1480
1493


Unsigned but KYTC-maintained:


1452


Incorrect endpoint locations (most are small adjustments):


1436
1439
1450
1456
1457
1461
1471
1475
1476
1479
1483
1499


Incorrect waypoint locations/missing waypoints (including shaping points to convert to waypoints):


1412
1414
1426
1428
1428Pai
1429
1453
1459
1460
1463
1469
1471
1473
1482
1491
1494
1496
1499


Incorrect endpoint/waypoint names:


1401
1411
1414
1415
1418
1426
1428
1439
1441
1442
1446
1448
1450
1453
1458
1460
1464
1465
1470
1472
1482
1486
1489
1492
1494
1496
1498


Reverse order to match mileposts:


1421
1491


All others appear to be in good shape.

KY1460 - points need moving onto the highway (and on interchanging roads); MayoTr_N -> KY3495; hidden point moved to nearby road?

These changes are included above.

[mapcat]

KY1486 has a connector to KY17 (officially called KY1486C) marked as KY1486, which leads to a strange 3-way intersection where every direction is KY1486. I'm inclined to leave it out, but am interested in hearing reasons for including it as well. There's at least one connector in the state highway system marked with a "C" sign, which tells me that if it were important enough, 1486C would have one as well (assuming they could fit one onto the sign).

[yakra]

KY1486 has a connector to KY17 (officially called KY1486C) marked as KY1486, which leads to a strange 3-way intersection where every direction is KY1486. I'm inclined to leave it out, but am interested in hearing reasons for including it as well. There's at least one connector in the state highway system marked with a "C" sign, which tells me that if it were important enough, 1486C would have one as well (assuming they could fit one onto the sign).

I see it colored green on the Kenton county SPRS map. (No caption though; too small/short.) So that means it's part of the state system, I think. (Right?) That combined with the signage would make me lean toward including it if I were drafting USAKY3. Signed as such, I'd probably choose KY1486 + a City suffix over calling it KY1486C, whatever it may be designated internally. Though, what to pick for said suffix? I'm glad I have the luxury of not having to ponder this too deeply. I don't know, man, I didn't do it. My feelings aren't terribly strong in any particular direction.

[mapcat]

@yakra: if you don't think it's too short, then I'll have a look-see how many of these there are out there of at least approximately the same length and bring in the whole lot of them at once.


Speaking of bringing in the whole lot: since usaky3 was complete when I inherited it and I conducted what I believe to be a thorough evaluation of all nearly 500 routes, making necessary corrections, can we promote this set to Active? 

[mapcat]

Any further review needed before activating the set?

[yakra]

Speaking of bringing in the whole lot: since usaky3 was complete when I inherited it and I conducted what I believe to be a thorough evaluation of all nearly 500 routes, making necessary corrections, can we promote this set to Active? 

Personally, I'd be okay with that. Well... on further thought, I'm waffling back & forth a bit...
On the other hand, as a new collaborator, a little checking wouldn't hurt, to make sure conventions are followed, point labels and positioning are A-OK, and the like.
On the third hand, for this reason, I kept a pretty close eye on your commits, and saw very few problems. I pointed out a few things, and you either fixed them or convinced me that your approach was correct. So yeah -- we're probably OK here, though another pair of eyes wouldn't hurt.

However, if you're going to go ahead with Operation Bring In the Whole Lot as described in your first paragraph, maybe it's better to hold off till that's done.
Anyone else have any Ayes or Nays, comments?