Texas Real Estate Detailed Listing Analysis - SAMPLE

**Sample:** 100 rows
**Total Rows:** ~14,857

**Overview**

This dataset delivers a curated, analytics-ready snapshot of **14,857 residential real-estate listings across Texas**, frozen in July 2025. Each row represents a unique property, with duplicates removed and **21 structured variables** covering architectural characteristics, pricing, lot attributes, listing status, and rich neighborhood-school metadata. The dataset also includes a fully sanitized listing narrative suitable for text-mining and natural-language modeling.

The free-text marketing descriptions have been cleaned of personally identifiable information—including emails, phone numbers, and street identifiers—while preserving linguistic nuance essential for NLP, valuation modeling, and explainability research. With strong completeness across core property and school-context attributes, this dataset offers a powerful foundation for AVMs, pricing analytics, localized demand modeling, and education-impact research within the Texas housing market.

**Provenance**

The dataset originates entirely from publicly accessible Texas real-estate listing sources and open government school directories. All records were collected and cleaned exclusively by **Maths with Kanchana LLC (Wyoming)**, led by **Kanchana Karunarathna**, a Kaggle Datasets Grandmaster and world-ranked contributor.

All personally identifiable information and agent/seller identifiers were removed or anonymized using deterministic regex filters and structured substitution methods. No authenticated APIs, paywalled feeds, or private data sources were used. The dataset is compiled ethically and adheres to GDPR-inspired standards of minimization and anonymization.

**Acknowledgements:**
Curated and engineered by Maths with Kanchana LLC (Kanchana Karunarathna).

**Use Cases**

This dataset enables high-value analytics and ML workflows on Databricks:

* **Automated Valuation Models (AVMs):**
Predict listPrice, price appreciation, or probability of sale using structured variables and NLP embeddings.

* **Education-Impact Pricing Models:**
Quantify how school quality, distance, and type influence property valuations across regions.

* **Time-on-Market & Conversion Modeling:**
Use listing status and soldOn to perform survival analysis and seasonality modeling.

* **Geospatial Demand Clustering:**
Leverage ZIP codes, construction era, lot size, and school density to identify growth corridors and supply gaps.

* **NLP-Driven Market Insights:**
Apply transformer models to redacted listing narratives to uncover premium-driving language patterns using SHAP/LIME.

**Column Dictionary**

* **type** (string): Primary property class (single_family, condo, etc.).
* **sub_type** (string): Detailed sub-classification (townhome, manufactured, etc.); sparsely populated.
* **text** (string): PII-redacted listing description suitable for NLP.
* **year_built** (decimal): Construction year; missing for some undeveloped lots.
* **soldOn** (date): Sale closing date where available.
* **lot_sqft** (decimal): Lot size in square feet.
* **baths** (decimal): Total bath count.
* **baths_full** (decimal): Count of full baths.
* **baths_full_calc** (decimal): MLS-calculated full-bath equivalent.
* **beds** (decimal): Number of bedrooms.
* **garage** (decimal): Garage capacity (stalls).
* **listPrice** (decimal): Asking price in USD at data freeze.
* **zip** (string): Texas ZIP code (100% coverage).
* **status** (string): Listing lifecycle stage (for_sale, pending, sold).
* **nearest_school_name** (string): Name of the nearest public/private school.
* **nearest_school_distance** (decimal): Distance to the nearest school (miles).
* **nearest_school_level** (string): School level (elementary, middle, high).
* **nearest_school_type** (string): Public/private/charter.
* **nearest_school_rating** (decimal): Rating (0–10) of nearest school.
* **avg_school_rating** (decimal): Average rating of schools in the ZIP.
* **school_count** (integer): Number of rated schools within the ZIP.