metadata
license: mit
configs:
- config_name: default
sep: ;
features:
- name: text
dtype: string
- name: label
class_label:
names:
- Airspaces
- Communication_and_Radar
- GPS
- Landing_Navaids
- Obstacles
- Parking
- Procedure
- Runway
- Services_and_Facilities
- Signs_and_Lights
- Taxiway
- Terminal_or_Enroute_Navaids
- Wildlife
task_categories:
- text-classification
language:
- en
size_categories:
- 1K<n<10K
Welcome to our NOTAMs dataset
This is a toy case for NLP classification but based on real published data. The goal is to classifiy a NOTAM (notice for airmen) based on its main subject. The dataset only contains NOTAMs in english. See more on NOTAM. You can find questions about this domain in fixingnotams.org.
Here are our definitions
| Names | Définitions | Ex of potential criticism, as classification is subjective ! |
|---|---|---|
| Airspaces | This NOTAM concerns an airspace definition for restriction or for authorization (including Airway) | |
| Communication_and_Radar | This NOTAM concerns communications (radio VHF, datalink CPDLC, ATIS, TWR, ATS) or radar or information provided by radio or digital way such as meteo information (ex : METAR) | Why perform a discrepancy between the data (METAR) and the system which provides it (WDI) ? To avoid to merge data & system. |
| GPS | This NOTAM concerns GPS, either for the area, either for the GPS augmentation system for landing. The procedures as GLS shall be classified as Procedure. | |
| Landing_Navaids | This NOTAM concerns landing navaids, mainly ILS (Instrument Landing System) or parts of ILS : LOC (lateral) or GLIDE (vertical), including Outer/Middle/Inner markers. This DOES NOT include the DME associated to the ILS, except if clearly explained in message (ILS, LDG, RWY). | why do not include DME ILS ? Data not obvious in message, sometimes available in Qcode, sometimes only in navaid ident |
| Obstacles | This NOTAM concerns obstacles (cranes, trees, vessels, …) | |
| Wildlife | This NOTAM concerns the wildlife issue, like birds hazard or any other animal hazard | |
| Parking | This NOTAM concerns the area where the aircrafts are parked. Usual names : apron, ramp, stand, parking. The helipads are classified as parking. | |
| Procedure | This NOTAM concerns procedures, either the standard published ones, either the specific applicable ones. If the NOTAM deals with AIP and has no obvious classification in another category, this NOTAM should be classified in this category. | |
| Runway | This NOTAM concerns runway state (closure, degraded surface, limited distance, runway arrest concern, work in progress around runway) . Runway signs and lights are NOT part of this category, except if the consequence is one of the above cases. | Why do not include everything abouta specific runway (RVRT, lights, marks...) ? To avoid to merge all in runway. |
| Services_and_Facilities | This NOTAM concerns services or facilities (fuel, meteo systems (ex : AWOS, RVR systems, windsock...), rescue, wash, pushback …), but not the signs or the lights. If the NOTAM deals with this subject and does not lead to close a precise part of the aerodrome, it should be classified in this category, rather than in another one. | Why perform a discrepancy between the data (METAR) and the system which provides it (WDI) ? To avoid to merge data & system. |
| Signs_and_Lights | This NOTAM concerns signs or lights. If the NOTAM deals with this subject and does not lead to close a part of the aerodrome, it should be classified in this category, rather than in another one. | Why do not include everything abouta specific runway (RVRT, lights, marks...) ? To avoid to merge runway and data. |
| Taxiway | This NOTAM concerns taxiway state or taxilane state (closure, degraded surface, work in progress around taxiway). Taxiway signs and lights are NOT part of this category, except if the consequence is one of the above cases. | Why do not include everything abouta specific runway (RVRT, lights, marks...) ? To avoid to merge all in taxiway. |
Feel free to share your comments
"To err is human. But AI makes mistakes, so AI is human." F.Hoofd 2025