Update README.md

README.md

@@ -1,66 +1,81 @@
-UFO Sightings – Cleaned & Unified Dataset (~327k rows)
+Perfect — now that we know the **exact real row count (~327,000 lines)** and the **actual schema** (the fields you pasted), here is the **correct, updated, minimal, technical HuggingFace README** for a *single JSONL file with ~327k entries*.
 
-This dataset merges several publicly available UFO sighting datasets from Kaggle into one cleaned, standardized, and enriched file.
-The goal is simply to provide a consolidated dataset instead of many fragmented sources with inconsistent formatting.
-
-This release contains a single JSONL file with approximately 327,000 records.
+No extra files.
+No embeddings.
+No UMAP description.
+No clustering method explanation unless the file already contains cluster_id & prob (which it does) — so we keep it minimal and factual.
 
-No private or identifying information was present in the original data.
+This README matches **exactly what you are providing**, nothing more.
 
-📦 Source
+Copy/paste this into your HuggingFace Space or Dataset README.
 
-All entries originate from publicly available UFO sighting datasets on Kaggle.
-Each row corresponds to a single reported sighting.
+---
 
-🧹 Cleaning / Normalization Performed
+# **UFO Sightings – Cleaned & Unified Dataset (~327k rows)**
 
-All rows in this unified file were standardized using the same basic rules:
+This dataset merges several publicly available UFO sighting datasets from Kaggle into one **cleaned**, **standardized**, and **enriched** file.
+The goal is simply to provide a consolidated dataset instead of many fragmented sources with inconsistent formatting.
 
-timestamps parsed and converted into a consistent t_utc (ISO-8601, UTC)
+This release contains **a single JSONL file** with approximately **327,000 records**.
 
-city/state/country fields harmonized where possible
+No private or identifying information was present in the original data.
 
-latitude/longitude coerced to floats
+---
 
-basic HTML/unicode cleanup in free-text descriptions (text)
+## **📦 Source**
 
-invalid or fully unparseable rows removed
+All entries originate from publicly available UFO sighting datasets on **Kaggle**.
+Each row corresponds to a single reported sighting.
 
-source field preserved as src
+---
 
-No interpretation or filtering based on content was performed.
+## **🧹 Cleaning / Normalization Performed**
 
-✨ Added Contextual Fields
+All rows in this unified file were standardized using the same basic rules:
 
-A small number of lightweight “sidecar” fields were added based on timestamp + coordinates:
+* timestamps parsed and converted into a consistent `t_utc` (ISO-8601, UTC)
+* city/state/country fields harmonized where possible
+* latitude/longitude coerced to floats
+* basic HTML/unicode cleanup in free-text descriptions (`text`)
+* invalid or fully unparseable rows removed
+* source field preserved as `src`
 
-moon_illum — moon illumination fraction
+No interpretation or filtering based on content was performed.
 
-moon_alt_deg — moon altitude in degrees
+---
 
-nearest_airport_code — closest airport (ICAO)
+## **✨ Added Contextual Fields**
 
-nearest_airport_km — distance to that airport in km
+A small number of lightweight “sidecar” fields were added based on timestamp + coordinates:
 
-wx_bucket — rough weather bucket (coarse category)
+* **`moon_illum`** — moon illumination fraction
+* **`moon_alt_deg`** — moon altitude in degrees
+* **`nearest_airport_code`** — closest airport (ICAO)
+* **`nearest_airport_km`** — distance to that airport in km
+* **`wx_bucket`** — rough weather bucket (coarse category)
 
 These values are approximate and should be treated as exploratory metadata only.
 
-🧩 Clustering Fields (Included in the File)
+---
 
-The dataset includes two fields that come from text-similarity grouping:
+## **🧩 Clustering Fields (Included in the File)**
 
-cluster_id — numeric label
+The dataset includes two fields that come from text-similarity grouping:
 
-prob — membership confidence
+* **`cluster_id`** — numeric label
+* **`prob`** — membership confidence
 
-These reflect text similarity, not verified categories or event types.
+These reflect **text similarity**, not verified categories or event types.
 They are included because they were already part of the cleaned file.
 
-📝 Field Reference
+---
+
+## **📝 Field Reference**
 
 Each row has the following structure (example):
 
+```
+{
   "uid": "scrubbed/row327047",
   "t_utc": "2013-09-09T09:51:00.000Z",
   "lat": 32.7152778,
@@ -80,36 +95,43 @@ Each row has the following structure (example):
   "nearest_airport_code": "KSAN",
   "reports_z": null,
   "wx_bucket": "unknown"
+}
+```
+
+### **Field descriptions**
+
+| Field                      | Type       | Notes                                            |
+| -------------------------- | ---------- | ------------------------------------------------ |
+| `uid`                      | string     | Stable row identifier                            |
+| `t_utc`                    | string     | Event timestamp, ISO-8601 UTC                    |
+| `lat`, `lon`               | float      | Approximate coordinates                          |
+| `city`, `state`, `country` | string     | Cleaned location fields (best-effort)            |
+| `text`                     | string     | Free-text sighting description                   |
+| `src`                      | string     | Original Kaggle dataset source                   |
+| `cluster_id`               | int        | Text-similarity cluster (for research use only)  |
+| `prob`                     | float      | Cluster membership probability                   |
+| `moon_illum`               | float      | Moon illumination (0–1)                          |
+| `moon_alt_deg`             | float      | Moon altitude in degrees                         |
+| `nearest_airport_km`       | float      | Distance to nearest airport                      |
+| `nearest_airport_code`     | string     | ICAO code                                        |
+| `wx_bucket`                | string     | Approximate weather category                     |
+| `reports_z`                | float/null | Unused placeholder field (kept for completeness) |
+
+---
+
+## **⚠️ Notes & Limitations**
+
+* Accuracy of timestamps and locations depends entirely on original reporting.
+* Weather buckets are coarse (not NOAA-grade).
+* Airport distances are approximate nearest-neighbor lookups.
+* Cluster labels are based solely on text similarity and do **not** reflect event reality.
+* No claims are made about the nature or validity of any sighting.
+
+---
+
+## **📄 License**
+
+Source data was public on Kaggle.
+This cleaned, merged, and lightly enriched version is released for **research and educational use**.
+Users should follow the original dataset licensing terms.
 
-
-Field descriptions
-
-Field	Type	Notes
-
-uid	string	Stable row identifier
-
-t_utc	string	Event timestamp, ISO-8601 UTC
-
-lat, lon	float	Approximate coordinates
-
-city, state, country	string	Cleaned location fields (best-effort)
-
-text	string	Free-text sighting description
-
-src	string	Original Kaggle dataset source
-
-cluster_id	int	Text-similarity cluster (for research use only)
-
-prob	float	Cluster membership probability
-
-moon_illum	float	Moon illumination (0–1)
-
-moon_alt_deg	float	Moon altitude in degrees
-
-nearest_airport_km	float	Distance to nearest airport
-
-nearest_airport_code	string	ICAO code
-
-wx_bucket	string	Approximate weather category
-
-reports_z	float/null	Unused placeholder field (kept for completeness)