---
language: en
license: mit
tags:
  - seo
  - local-seo
  - nap
  - consistency
---

# NAP-Consistency-Checker: Local SEO NAP and Entity Validation

**Type:** Commercial | **Domain:** SEO, Local SEO  
**Hugging Face:** [syeedalireza/nap-consistency-checker](https://huggingface.co/syeedalireza/nap-consistency-checker)

Validate NAP (Name, Address, Phone) consistency and entity alignment across pages and citation sources.

## Author

**Alireza Aminzadeh**  
- Hugging Face: [syeedalireza](https://huggingface.co/syeedalireza)  
- LinkedIn: [alirezaaminzadeh](https://www.linkedin.com/in/alirezaaminzadeh)  
- Email: alireza.aminzadeh@hotmail.com

## Problem

Local SEO depends on consistent NAP across site and citations. Detecting variations and normalizing entities reduces confusion for users and search engines.

## Approach

- **Input:** Text snippets or structured NAP fields (name, address, phone) from multiple sources.
- **Output:** Consistency flags (match/mismatch), normalized canonical form, optional entity resolution (same business or not).
- **Models:** Rule-based normalization (address/phone parsing) + optional transformer/NER for extracting NAP from raw text; embedding similarity for entity matching.

## Tech Stack

| Category | Tools |
|----------|------|
| NLP | Hugging Face Transformers, sentence-transformers (optional) |
| Rules | regex, standard phone/address parsing |
| Data | pandas, NumPy |

## Setup

```bash
pip install -r requirements.txt
```

## Usage

```bash
python inference.py --input data/citations.csv --reference data/reference_nap.csv --output output/consistency_report.csv
```

- `normalize.py` provides NAP normalization helpers (used by inference). No separate “build canonical” step; the reference is read from the reference CSV.

## Project structure

```
09_nap-consistency-checker/
├── config.py
├── normalize.py       # NAP normalization helpers
├── inference.py      # Compare citations to reference NAP
├── requirements.txt
├── .env.example
├── data/
│   ├── reference_nap.csv   # Canonical NAP (one row: name, address, phone)
│   └── citations.csv       # Source, name, address, phone per citation
└── output/                 # consistency_report.csv
```

## Data

- **Sample data (included):** `data/reference_nap.csv` (one row: `name`, `address`, `phone`), `data/citations.csv` (columns: `source`, `name`, `address`, `phone`). Output: `name_match`, `address_match`, `phone_match`, `nap_consistent`.

## License

MIT.