Document centreline & centreline_intersection pitfalls and create useful transformations

[radumas]

The centreline contains more than just streets. We should document pitfalls to avoid when querying it and also create useful transformations (in particular of the intersections table).

[raha1986]

I am using current issue to add all the information I learnt about gis.centreline, and gis.centreline_intersections. Additional to the datasets, I am adding information about the analysis that has been done (I was involved) using these datasets. Hopefully, after a while, it will be re-organised into a useful wiki about these dataset, and finally, a source to organise metadata.

I start with the general information, some parts I leave blank (...). It would be helpful to fill them during the time. I will add some information day by day and perhaps, each purge, it can be reorganised!

gis.centreline

---

The gis.centreline is a Linear feature representing the TO Road Network.

• [ ] It's collected by ...
• [ ] It's maintained by ...

Online version

• [ ] A link for public to access is under OpenData (updated on ... )
• [ ] Regular updates (online version) : Quarterly
• [ ] recent update (online version) : June 2022

Our on server version

• [ ] The physical address for us is under gis schema in bigdata database installed on EC2 server
• [ ] What's current version's update date?

UPDATES/ CHANGES

The changes between different versions are not very clear! There has been some attempts to compare them, however it's on hold due to ... . There is no current methodology to compare the most recent centreline layer with the former one. The changes ranging from deletion of a segment, geometric changes in a part of a segment or values of its attributes to appearing new segments or new attributes for segments.

ATTRIBUTES

1- fcode_desc

Unique values for fcode_desc in gis.centreline:

• "Expressway"

• "Major Arterial"

• "Minor Arterial"

• "Collector"

• "Local": If looking for "local" street, we choose only this type. Example issue 260

• "Other": First inspections, and then include

• "Pending": First inspections, and then include

• "Expressway Ramp"

• "Major Arterial Ramp"

• "Minor Arterial Ramp"

• "Collector Ramp"

• "Other Ramp"

• "Laneway"

• "Minor Railway"

• "Access Road"

• "Major Railway"

• "Walkway"

• "Minor Shoreline (Land locked)"

• "Geostatistical line"

• "Trail"

• "Ferry Route"

• "River"

• "Major Shoreline"

• "Hydro Line"

• "Creek/Tributary"

• "Busway"

gis.centreline_intersections

---

ATTRIBUTES

classifi7

Values for classifi7 in gis.centreline_intersection

values for elavatio10 can be used to identify Major intersection or locals

To the extend I realised both Major and Minor values for elavatio10 can be used for local and major intersections.

DISCUSSION: I am going to add some examples and then we can discuss

• For "Major" intersections: ('Major', 'Minor', 'Expressway')
• For "Local" intersections: ('Major', 'Minor')

Analysis related to intersections/ centreline

---

1.Collisions at intersections

• a 20m buffer around an intersection
• Use collisions_replicator.events to find relevant collisions to intersections

2.Collisions on segments/ midblocks

• a 20m buffer around a segment
• Use collisions_replicator.events to find relevant collisions to intersections

[raha1986]

The gis.centreline was updated at least more than 3 years ago.

[raha1986]

#	Field_name	Details
0	gid
1	geo_id
2	lfn_id	line   feature name id
3	lf_name	line feature name
4	address_l
5	address_r
6	oe_flag_l	Even/Odd   address numbers on left side of street
7	oe_flag_r	Even/Odd   address numbers on right side of street
8	lonuml	Lowest address   number for left side of the street segment (LO_NUM_L)
9	hinuml	Highest   address number for left side of the street segment (HI_NUM_L)
10	lonumr	Lowest address   number for right side of the street segment (LO_NUM_L)
11	hinumr	Highest   address number for right side of the street segment (HI_NUM_L)
12	fnode	from_node
13	tnode	to_node
14	fcode	Feature   code
15	fcode_desc	Feature Descriptions
16	juris_code	Jurisdiction   Code
17	objectid

[raha1986]

Based on limited provided information , the centreline dataset is a geographic representation of several Linear features such as transportation corridors, administrative boundaries, and shorelines. So, I was wondering if in future, it makes sense to have a subset dataset that is a representation of transportation corridors and roads. So we would work with a smaller dataset and exclude some features like 'geostatistical Lines' or 'Hydro Lines` from it.

`

fcode	fcode_desc	count	descriptions
201100	Expressway	1452
201101	Expressway Ramp	1265
201200	Major   Arterial	6067
201201	Major Arterial Ramp	199
201300	Minor   Arterial	3575
201301	Minor Arterial Ramp	3
201400	Collector	6397
201401	Collector Ramp	17
201500	Local	24740
201600	Other	2023
201601	Other Ramp	12
201700	Laneway	4132
201800	Pending	846
201801	Busway	8
201803	Access Road	41
202001	Major Railway	831
202002	Minor   Railway	16
203001	River	1013
203002	Creek/Tributary	20	Creek is a stream, brook, or minor tributary of a river. A tributary, or affluent, is a stream or river that flows into a larger stream or main stem river or a lake.
204001	Trail	9601
204002		332
205001	Hydro Line		named as "Hepc" in `lf_name` --> Hydro-Electric Power Commission of Ontario
206001	Major Shoreline	267
206002	Minor Shoreline (Land locked)	2
207001	Geostatistical line	4228	Imaginary boundaries like planning, municipals
208001	Ferry Route	19

`

[chmnata]

FYI, these are the content of the centreline intersection layer from GCC view

INT_ID = INTERSECTION_ID (Unique id of the intersection) ELEV_ID = ELEVATION_ID (Unique id of the elevation) INTERSEC5 = INTERSECTION_DESC (Intersection Description: two cross streets) CLASSIFI6 = CLASSIFICATION (Intersection classification) CLASSIFI7 = CLASSIFICATION_DESC (Intersection classification description) NUM_ELEV = NUMBER_OF_ELEVATIONS (Number of different elevation in the intersection) ELEVATIO9 = ELEVATION_FEATURE_CODE (Feature code of the elevation) ELEVATIO10 = ELEVATION_FEATURE_CODE_DESC (Feature code description of the elevation) ELEV_LEVEL = ELEVATION_LEVEL (Elevation level number) ELEVATION = ELEVATION (Elevation height) ELEVATIO13 = ELEVATION_UNIT (Elevation measurement unit ) HEIGHT_R14 = HEIGHT_RESTRICTION (Height restriction) HEIGHT_R15 = HEIGHT_RESTRICTION_UNIT (Height restriction unit) X = X (Easting, in MTM NAD 27(3 degree) Projection) Y = Y (Northing, in MTM NAD 27(3 degree) Projection) LONGITUDE = LONGITUDE (Longitude in WGS84 Coordinate System Easting of address point, GCS Projection) LATITUDE = LATITUDE (Latitude in WGS84 Coordinate System) OBJECTID = OBJECTID (Unique system identifier)

[raha1986]

Some points about centreline_intersection:

1- The int_id is not unique! There are some cases where different roads or linear features, in general, are passing over each other. Therefore, the point is represented by a single node and an identical int_id.

2- One way of differentiating the intersections and perhaps linked centreline, is using the elevatio10.

3. Be aware that not all of the elevatio10 are related to transportation. "Minor", "River", "Railway", "Pedatraian", "No Intersection", "Pseudo" "Error No Domain Descrption","Expressway", "Unknown", "Laneway", "Major", "Cul de sac", "Utility"

4. The ones I think are relevant are "Pedatraian", "Minor", "Expressway", "Laneway", "Major", "Cul de sac"

QUESTIONs/ curiosity

1. I am not sure about the "pseudo"! int_id=13441579
2. It might be also useful to know about elev_level. For example, whether elev_level=0 means ground level or not, or what would be elev_level=9

[raha1986]

Around 753 intersections, do not have names (intersec5). It's not a huge proportion, less than 2% of the intersections but distributed over the city. Just good to know. A good proportion of it is along the lake, those that I think are less important, especially for collisions.

Mostly not intersection and unknown, then there are pseudo, and cul de sac ones. Around 5 of the Minor intersections are also without names.

[mkewins]

@raha1986 our former developer, Evan, did a lot of work to explore and interpret the centreline for MOVE (I can forward you his documentation, if you're interested). Some of what you've noted here impacts MOVE as well, for example, segments without names. There are downstream effects on our users – things they find challenging or frustrating. Flagging this because, if you make progress or decisions on how to interpret and transform the centreline, we'd love to use your learnings (or even transformations).