aerostratospheric/msds-open-models
Tabular Classification • Updated • 50
Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
schema: string
date_utc: timestamp[s]
generated_at_utc: timestamp[s]
curator: string
disclaimer: string
methods: struct<absolute_thresholds: struct<temperature_c: struct<alert_high: int64, watch_high: int64, watch (... 363 chars omitted)
child 0, absolute_thresholds: struct<temperature_c: struct<alert_high: int64, watch_high: int64, watch_low: int64, alert_low: int6 (... 197 chars omitted)
child 0, temperature_c: struct<alert_high: int64, watch_high: int64, watch_low: int64, alert_low: int64>
child 0, alert_high: int64
child 1, watch_high: int64
child 2, watch_low: int64
child 3, alert_low: int64
child 1, pressure_msl_hpa: struct<watch_low: int64, info_high: int64>
child 0, watch_low: int64
child 1, info_high: int64
child 2, wind_speed_ms: struct<watch: int64, alert: int64>
child 0, watch: int64
child 1, alert: int64
child 3, wave_height_m: struct<watch: int64, alert: int64>
child 0, watch: int64
child 1, alert: int64
child 4, docs: string
child 5, config: string
child 1, zscore_normalization: struct<z_pop: string, z_sample: string, z_mad: string, s_range: string, percentile: string>
child 0, z_pop: string
child 1, z_sample: string
child 2, z_mad: string
child 3, s_range: string
child 4, percentile: string
child 2, examples_doc: string
counts: struct<total: int64, alert: int64, watch: int64, info: int64>
...
percentile: double
child 7, mean: double
child 8, median: double
child 9, pstdev: double
child 10, sample_stdev: double
child 11, mad: double
zscore_by_city: list<item: struct<id: string, name: string, methods: struct<n: int64, x: double, z_pop: double, z_sa (... 148 chars omitted)
child 0, item: struct<id: string, name: string, methods: struct<n: int64, x: double, z_pop: double, z_sample: doubl (... 136 chars omitted)
child 0, id: string
child 1, name: string
child 2, methods: struct<n: int64, x: double, z_pop: double, z_sample: double, z_mad: double, s_range: double, percent (... 93 chars omitted)
child 0, n: int64
child 1, x: double
child 2, z_pop: double
child 3, z_sample: double
child 4, z_mad: double
child 5, s_range: double
child 6, percentile: double
child 7, mean: double
child 8, median: double
child 9, pstdev: double
child 10, sample_stdev: double
child 11, mad: double
inputs: struct<global_cities: bool, ndbc: bool, casey: bool, baseline_files_used: int64>
child 0, global_cities: bool
child 1, ndbc: bool
child 2, casey: bool
child 3, baseline_files_used: int64
anomaly_file_count: int64
archive_by_date: list<item: timestamp[s]>
child 0, item: timestamp[s]
anomaly_files_sample: list<item: string>
child 0, item: string
role: string
git_policy: string
to
{'generated_at_utc': Value('timestamp[s]'), 'anomaly_file_count': Value('int64'), 'anomaly_files_sample': List(Value('string')), 'archive_by_date': List(Value('timestamp[s]')), 'role': Value('string'), 'git_policy': Value('string')}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
schema: string
date_utc: timestamp[s]
generated_at_utc: timestamp[s]
curator: string
disclaimer: string
methods: struct<absolute_thresholds: struct<temperature_c: struct<alert_high: int64, watch_high: int64, watch (... 363 chars omitted)
child 0, absolute_thresholds: struct<temperature_c: struct<alert_high: int64, watch_high: int64, watch_low: int64, alert_low: int6 (... 197 chars omitted)
child 0, temperature_c: struct<alert_high: int64, watch_high: int64, watch_low: int64, alert_low: int64>
child 0, alert_high: int64
child 1, watch_high: int64
child 2, watch_low: int64
child 3, alert_low: int64
child 1, pressure_msl_hpa: struct<watch_low: int64, info_high: int64>
child 0, watch_low: int64
child 1, info_high: int64
child 2, wind_speed_ms: struct<watch: int64, alert: int64>
child 0, watch: int64
child 1, alert: int64
child 3, wave_height_m: struct<watch: int64, alert: int64>
child 0, watch: int64
child 1, alert: int64
child 4, docs: string
child 5, config: string
child 1, zscore_normalization: struct<z_pop: string, z_sample: string, z_mad: string, s_range: string, percentile: string>
child 0, z_pop: string
child 1, z_sample: string
child 2, z_mad: string
child 3, s_range: string
child 4, percentile: string
child 2, examples_doc: string
counts: struct<total: int64, alert: int64, watch: int64, info: int64>
...
percentile: double
child 7, mean: double
child 8, median: double
child 9, pstdev: double
child 10, sample_stdev: double
child 11, mad: double
zscore_by_city: list<item: struct<id: string, name: string, methods: struct<n: int64, x: double, z_pop: double, z_sa (... 148 chars omitted)
child 0, item: struct<id: string, name: string, methods: struct<n: int64, x: double, z_pop: double, z_sample: doubl (... 136 chars omitted)
child 0, id: string
child 1, name: string
child 2, methods: struct<n: int64, x: double, z_pop: double, z_sample: double, z_mad: double, s_range: double, percent (... 93 chars omitted)
child 0, n: int64
child 1, x: double
child 2, z_pop: double
child 3, z_sample: double
child 4, z_mad: double
child 5, s_range: double
child 6, percentile: double
child 7, mean: double
child 8, median: double
child 9, pstdev: double
child 10, sample_stdev: double
child 11, mad: double
inputs: struct<global_cities: bool, ndbc: bool, casey: bool, baseline_files_used: int64>
child 0, global_cities: bool
child 1, ndbc: bool
child 2, casey: bool
child 3, baseline_files_used: int64
anomaly_file_count: int64
archive_by_date: list<item: timestamp[s]>
child 0, item: timestamp[s]
anomaly_files_sample: list<item: string>
child 0, item: string
role: string
git_policy: string
to
{'generated_at_utc': Value('timestamp[s]'), 'anomaly_file_count': Value('int64'), 'anomaly_files_sample': List(Value('string')), 'archive_by_date': List(Value('timestamp[s]')), 'role': Value('string'), 'git_policy': Value('string')}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Aerostratospheric Defense GIR
A plain-language guide to our open Geospatial Information Repository
In one sentence: We automatically collect free, public map and hazard information, organize it, and save dated copies here — without secret military data.
Live web page: Defense GIR on midwestsds.com · Charts: docs/charts/ · Status JSON: data/status/ · Latest report: reports/latest.md · Data packages: docs/DATA_PACKAGES.md
What is this, in plain English?
Imagine you need a shared notebook of public facts about the world—storms that were already warned about, earthquakes that already happened, free satellite “library cards,” public cyber patch lists, and open research balloon flight summaries. You do not want secrets. You want something you can audit, re-run, and explain to a non-specialist.
That notebook is this repository.
GIR means Geospatial Information Repository: place-based information stored in an organized library. Software here downloads free public feeds on a schedule, checks whether the download worked, keeps a short history in git, draws simple charts, publishes a US open-status summary for the website banner, and writes a daily executive summary. Partner or restricted sensor products from Aerostratospheric platforms (when fielded) stay out of this public tree on purpose.
Think of it as the weather-and-world briefing binder that sits beside more sensitive tools—not the classified mission folder, not a weapons system, and not a substitute for official emergency alerts on your phone.
| Word | Everyday meaning |
|---|---|
| Geospatial | Facts tied to a place on Earth |
| Information | Measurements and events (quakes, alerts, satellite catalogs, public cyber lists, flight log summaries, …) |
| Repository | An organized digital filing cabinet with history |
Daily data packages
Every successful automation run produces a coherent open-tier package:
| Area | Location |
|---|---|
| Manifest (ok/total per source) | data/manifests/manifest_latest.json |
| UOGW anomalies | data/anomalies/ |
| Hazards (EONET, USGS, NWS, DONKI) | data/events/ |
| Sentinel-2 STAC index | data/imagery_index/ |
| Defense-open samples (CISA KEV, OpenSky, OurAirports, …) | data/defense_open/ |
| US open-status banner | data/status/ |
| Charts (SVG + Mermaid) | docs/charts/, docs/GRAPHS.md |
| Executive summary | reports/daily/, reports/latest.md |
Full map: docs/DATA_PACKAGES.md · Feature catalog: docs/FEATURES.md
How it is used (sample use cases)
These examples are open-tier only. They describe realistic ways people and programs use this kind of public data—not targeting, not classified operations.
1. Morning open briefing for a small research team
A Midwest balloon or environmental team opens the GIR web page, checks the US open status strip (GREEN/YELLOW/…), skims the latest exec report, and notes whether any public research flight is marked active in the flight log.
2. “What can an uncleared partner already see?”
Before a partner discussion, staff review the Sentinel-2 index and public hazard layers. That answers a simple question in lay terms: which free satellite scenes and public alerts already exist for this region?
3. Classroom or STEM outreach
A teacher uses the earthquake magnitude chart and EONET categories to show how open science feeds work. Students learn that “defense-adjacent” public data can mean weather, disasters, and transparency lists—not secret bases.
4. Cyber hygiene desk check
An IT lead glances at the CISA KEV sample in data/defense_open/. It is a public “patch these known-exploited holes first” catalog.
5. Flight log continuity for public launches
When Aerostratospheric publishes a public balloon event summary, it is appended to data/flight_logs/.
6. Automation and reproducibility
python3 scripts/ingest_open_tier.py
python3 scripts/compute_us_open_status.py
python3 scripts/generate_gir_charts.py
python3 scripts/generate_daily_exec_summary.py
# or:
bash scripts/daily_gir_automation.sh
7. Grant / SAM.gov narrative support
Open GIR materials illustrate a responsible open-data posture: public inputs, clear tiers, disclaimers, and no claim to replace NWS, USGS, or military command systems.
Non-use cases (on purpose): sole life-safety alerting, classified basing lists, targeting folders, or kinetic system control.
Executive reports
Open-tier daily executive summaries are generated automatically by the same GitHub Actions pipeline.
| Resource | Link |
|---|---|
| Latest summary | reports/latest.md |
| Daily archive | reports/daily/ |
Weekly roll-ups can be added to the same pipeline later if needed.
Live visual charts
python3 scripts/generate_gir_charts.py
See docs/charts/ for the full category gallery.
Access tiers
flowchart TB
O["Open — this public repo"] --> P["Partner — agreement"] --> R["Restricted — authorized only"]
Related open ecosystem
| Repository | Role |
|---|---|
| Unified-Open-Global-Weather | Multi-layer atmospheric commons + anomaly report |
| msds-data | Casey ground weather + planned HAB flight packages |
| International-Ground-Data-Repository | IGRA / international ground indexes |
| x2griffon | Payload platform documentation |
What’s inside (short tour)
| Area | Plain meaning |
|---|---|
| Satellite / EO | Library-card indexes for free imagery (Sentinel, …) |
| Military-marked (public) | Keyword heuristics on public airport names — not official basing |
| Alerts & hazards | USGS, NWS, EONET, UOGW, DONKI |
| Cyber & spending | CISA KEV + USAspending samples |
| Flight logs | Public research balloon / test summaries |
| US open status | Banner levels from public feeds only — not DEFCON |
| Executive reports | Daily open-tier briefing summaries |
Automation
bash scripts/daily_gir_automation.sh
Schedule: 06:00 and 18:00 UTC (plus manual Actions dispatch).
Pipeline: open-tier ingest → US open-status → charts → daily executive summary → git commit.
Details: docs/DAILY_AUTOMATION.md
Repository hygiene
| Doc | Purpose |
|---|---|
| CONTRIBUTING.md | How to propose changes (open-tier only) |
| SECURITY.md | How to report vulnerabilities |
| CITATION.cff | Formal citation |
| docs/DATA_POLICY.md | What may / may not enter the tree |
| docs/DATA_PACKAGES.md | Package map for data consumers |
| docs/FEATURES.md | Full feature catalog |
Links
| Resource | URL |
|---|---|
| Defense GIR page | https://midwestsds.com/aerostratospheric-defense-gir.html |
| Defense Systems | https://midwestsds.com/aerostratospheric-defense-systems.html |
| Latest exec report | reports/latest.md |
| Daily report archive | reports/daily/ |
| Data packages guide | docs/DATA_PACKAGES.md |
| Contact | https://midwestsds.com/contact/index.php?a=add |
Aerostratospheric is a registered SAM.gov entity. Passive sensing · intelligence support · communications — no kinetic weapons. Open-tier data is not an official warning service alone.
- Downloads last month
- 2,168
Models trained or fine-tuned on aerostratospheric/gir