Hello Finnish OSM community,
I’d like to announce a new MapRoulette challenge: Finland - Nodes Without Primary Tags. The aim of this challenge is to improve data completeness by adding appropriate primary tags to OSM nodes that currently have a name=* tag but lack top-level tags defining the features (such as amenity=, shop=, place=*, etc.)
Your comments, feedback, and suggestions are very welcome.
Thank you,
Priscilla - TomTom
These “without primary tags” quests have popped up all over the place, so I’m just picking a random one to ask - what query is actually being used here?
OSM has a huge variety of what could be called “primary tags” (and there has been much debate over that), hence the question…
Thanks for raising the question!
These challenges are based on an Overpass query that selects nodes that have a name=* tag but lack any top‑level descriptive tag such as amenity=*, shop=*, tourism=*, place=*, office=*, and so on. In other words, these are named points that currently don’t indicate what the feature actually is. The goal is to help improve data completeness by adding an appropriate primary tag when it can be confidently determined.
More details in the respective challenge description:
Finland: MapRoulette
This challenge is part of the data‑quality initiatives run by TomTom within the OSM organised editing framework. You can find an overview of these activities here:
A few important notes for mappers: If the available information (imagery, local knowledge, website, signage, etc.) is insufficient to determine the correct primary tag, or if the feature is ambiguous, outdated, or no longer exists, contributors are encouraged to mark the task as “Can’t complete”. The intention is not to force tagging decisions where the data isn’t clear.
Happy to adjust the challenge if you see ways to improve it.
Is the actual query available anywhere?
I’m sharing below the SQL query we used on an offline data extract, which allowed us to run a few quality checks—such as filtering out some obvious false positives.
To make the process easier to follow, I’ve broken the workflow down into the following steps:
1. Selecting named nodes
We started by selecting all point features that contain a name=* tag and represent real OSM objects (i.e., not deleted or legacy IDs).
2. Excluding lifecycle, deprecated, or non‑current features
Nodes whose tags indicate demolished, abandoned, disused, proposed, construction, or otherwise non‑active states were excluded.
This also includes objects tagged with past‑tense markers (e.g., was:*) or temporary states.
3. Filtering out nodes where the only tags are secondary, technical, or non‑descriptive
We removed points that have tags which are known not to represent standalone mappable features.
Examples include:
structural or metadata‑like tags (e.g., level, location, indoor)
functional markers (e.g., craft=*, club=*, utility=*)
golf tagging elements, amusement‑ride components
piste, advertising, industrial, telecom, pipeline, marker tags
various temporary or prefix-form highway tags
ref-specific or operational-only tags
These tags indicate auxiliary data rather than a primary feature.
4. Restricting to nodes without any primary feature tag
We kept only the nodes without primary tags such as:
amenity, shop, landuse, railway, historic, tourism, building, natural, power, office, place, barrier, public_transport, and many more.
If a point had any recognized primary feature tag, it was excluded.
5. Excluding named nodes that likely represent numerical labels
Nodes whose names consist mainly of digits (e.g., streetlamp numbers, parking spot numbers) were removed unless they represent well‑known numeric patterns like “24/7”.
6. Spatial checks to remove nodes attached to linear features
To avoid picking up unnamed components of other mapped objects, we excluded nodes that physically intersect with:
highways
boundaries
railways
waterways
This prevents false positives like named bollards, mile markers, or boundary stones.
7. Removing nodes inside special landuse areas
Nodes located within polygons tagged as:
landuse=military
aeroway=aerodrome
were excluded, as names inside these areas frequently represent internal infrastructure not meant as standalone public features.
SELECT
-- p.*,
p.osm_id, p.name, CAST(p.tags -> 'osm_timestamp' AS DATE) AS last_edit_date,
ST_AsText(p.way) AS geom
FROM {schema}.planet_osm_point p
WHERE
p.name IS NOT NULL
AND NOT EXISTS (
SELECT 1
FROM each(p.tags) AS t(key, val) -- remove lifecycle prefixes
WHERE t.key LIKE ANY (ARRAY[
'%demolished%', '%disused%', '%abandoned%', '%ruins%', '%removed%', '%razed%', '%destroyed%',
'%was:%', '%:was%', '%proposed%', '%construction%', '%damage%', '%closed%', '%planned%',
'%historic:%', '%:historic%', '%seamark%', '%:disabled%', '%disabled:%', '%fire_hydrant%'
])
)
AND NOT EXISTS (
SELECT 1
FROM each(p.tags) AS t(key, value) -- remove FP features
WHERE key IN (
'area:highway', 'craft', 'club', 'entrance', 'healthcare', 'emergency', 'heritage',
'utility', 'mountain_pass', 'traffic_sign', 'beacon:type', 'fence_type',
'kerb', 'level', 'location', 'indoor', 'door', 'entrance', 'border_type',
'building:part', 'golf', 'disc_golf', 'tee', 'rough', 'bunker', 'hole',
'fairway', 'green', 'pin', 'out_of_bounds', 'par',
'slot_car', 'roller_coaster', 'dark_ride',
'attraction', 'shop', 'climbing', 'playground',
'piste:difficulty', 'piste:grooming', 'piste:type',
'advertising', 'sac_scale', 'mooring',
'usage', 'substance',
'allotments', 'cemetery', 'length', 'police', 'parking',
'ref:ProRailID', 'ref:ProRailSpoortak', 'temporary:highway',
'not:highway', 'old:highway', 'no:highway',
'bicycle_parking', 'horse_scale',
'removed:highway:2015', 'highway:2018', 'highway:-2019',
'highway:2022', 'highway:-2022', 'highway:-2023', 'geological', 'waterway:sign',
'information', 'university', 'telecom', 'industrial', 'animal', 'checkpoint', 'whitewater',
'community', 'incline', 'pipeline', 'valve', 'actuator', 'pedagogy', 'marker'
)
)
AND p.highway IS NULL
AND p.boundary IS NULL
AND p.layer IS NULL
AND p.amenity IS NULL
AND p.barrier IS NULL
AND p.ele IS NULL
AND p.railway IS NULL
AND p.leisure IS NULL
AND p.historic IS NULL
AND p.junction IS NULL
AND p.military IS NULL
AND p.office IS NULL
AND p.place IS NULL
AND p.power IS NULL
AND p.shop IS NULL
AND p.waterway IS NULL
AND p.aeroway IS NULL
AND p.man_made IS NULL
AND p.public_transport IS NULL
AND p.route IS NULL
AND p.aerialway IS NULL
AND p.sport IS NULL
AND p.building IS NULL
AND p.landuse IS NULL
AND p.tourism IS NULL
AND p."natural" IS NULL
AND p.disused IS NULL
AND p.osm_id > 0
AND NOT ((p.name ~ '[0-9].*[0-9]') -- remove values that have equal to or more than two digits in the name
AND p.name !~ '(24/7|247|365|24)')
AND NOT EXISTS ( -- remove features intersecting with highways
SELECT 1
FROM {schema}.planet_osm_line l
WHERE
ST_Intersects(p.way, l.way)
AND l.highway IS NOT NULL
)
AND NOT EXISTS ( -- remove features intersecting with boundaries
SELECT 1
FROM {schema}.planet_osm_line l
WHERE
ST_Intersects(p.way, l.way)
AND l.boundary IS NOT NULL
)
AND NOT EXISTS ( -- remove features intersecting with railways
SELECT 1
FROM {schema}.planet_osm_line l
WHERE
ST_Intersects(p.way, l.way)
AND l.railway IS NOT NULL
)
AND NOT EXISTS ( -- remove features intersecting with waterways
SELECT 1
FROM {schema}.planet_osm_line l
WHERE
ST_Intersects(p.way, l.way)
AND l.waterway IS NOT NULL
)
AND NOT EXISTS (
SELECT 1
FROM {schema}.planet_osm_polygon military
WHERE military.landuse = 'military'
AND ST_Intersects(p.way, military.way)
)
AND NOT EXISTS (
SELECT 1
FROM {schema}.planet_osm_polygon aerodrome
WHERE aerodrome.aeroway = 'aerodrome'
AND ST_Intersects(p.way, aerodrome.way)
)
Excluded nodes:
Excluded geometries: line, polygon, relation
Not quite related to this challenge, but for better or for worse, at least in Finland piste is sometimes used as the “main tag” on ways. E.g. nordic skiing tracks running over a lake or a commercially maintained snowshoe route.
Noted 
Thanks
Hi OSM Finland 
!
Just a quick, friendly nudge about the Finland: Nodes without primary tags MapRoulette challenge — we’re down to 465 remaining tasks!
These nodes already have a name=* tag and in most cases, also have contact details or website that would help to determine what the POI represents and add the actual primary tag. Confirming the POI type and adding the correct tags (amenity, shop, leisure, tourism, office, place, etc.) significantly improves data completeness and usability.
The tasks are straightforward: verify the feature using imagery, other tag info available or reliable sources, check for duplicates, add the appropriate tags, or mark the task accordingly if the node is outdated or already fixed.
If you have a moment, a bit of local knowledge can make a real difference here 
Any help or feedback is appreciated!