thaieda 1.0.1

AutoEDA สำหรับข้อมูลภาษาไทย — Exploratory data analysis that speaks Thai

ThaiEDA

Exploratory data analysis that actually understands Thai.

---

What is ThaiEDA?

ThaiEDA is a Python library that automates exploratory data analysis for Thai-language datasets. You give it a DataFrame, it gives you back a full report — column types, quality issues, anomalies, cross-column insights, charts, and an HTML report. All in one line.

It handles the things generic EDA tools miss: Buddhist Era dates, Thai numerals, zero-width spaces, mojibake encoding, Thai month names, and PII like phone numbers and national ID cards.

---

Quick Start

pip install thaieda

import thaieda
import pandas as pd

df = pd.read_csv("data.csv")
result = thaieda.run(df)          # that's it — full EDA in one line
result.to_html("report.html")     # self-contained HTML report

Want everything (Thai tokenizer, NER, ML, Excel, stats, LLM)?

pip install "thaieda[all]"

---

Why ThaiEDA?

Generic tools don't understand Thai data. Pandas Profiling, ydata-profiling, and Sweetviz are great — until you feed them Thai data. They miss Buddhist Era years (พ.ศ.), Thai numerals (๑๒๓), zero-width spaces that break tokenization, and mojibake from TIS-620 encoding. ThaiEDA catches all of these.

Privacy-first LLM analysis. Want to ask an LLM about your data but can't send raw rows to a cloud API? ThaiEDA has 4 privacy modes — the default sends zero raw data off your machine. Just stats and insights. Perfect for government, finance, and medical data under PDPA.

Insights, not just summaries. Most EDA tools show you df.describe() with nicer formatting. ThaiEDA has a cross-column insight engine that finds non-obvious patterns — "column A strongly predicts column B", "this group is 3× higher than average", "this column has outliers at row 47" — ranked by statistical interestingness with Benjamini-Hochberg correction.

One line to get everything. thaieda.run(df) chains the full pipeline: type detection → data cleaning → quality checks → anomaly detection → insight discovery → visualization → HTML report. No config needed.

---

How It Works

DataFrame
    │
    ▼
┌─────────────────────────────────────────┐
│  thaieda.run(df)                        │
│                                         │
│  1. detect    → column types             │
│  2. clean     → fix encoding/numerals/BE │
│  3. quality   → nulls, placeholders, BE │
│  4. anomaly   → statistical + text      │
│  5. insights  → 6 cross-column patterns │
│  6. viz       → auto charts (Thai font) │
│  7. report    → self-contained HTML     │
│                                         │
│  + optional: LLM analysis (4 modes)     │
└─────────────────────────────────────────┘
    │
    ▼
EDAResult
  .to_html()      → report.html
  .to_dict()      → Python dict
  .to_json()      → JSON string
  .insights       → insight cards
  .cleaned_df     → cleaned DataFrame
  .quality_issues → list of issues
  .anomalies      → anomaly findings
  .llm_response   → LLM analysis (if enabled)

---

Examples

One-Line EDA

import thaieda
import pandas as pd

df = pd.read_csv("data.csv")

# Full pipeline: detect → clean → quality → insights → viz → report
result = thaieda.run(df)

# Access results
result.to_html("report.html")
print(result.insights)           # cross-column insight cards
print(result.quality_issues)     # data quality findings
print(result.notes)              # pipeline notes/warnings

# Alias works too
result = thaieda.EDA(df)

With LLM Analysis (Privacy-Safe)

import thaieda

# Default: zero raw data leaves your machine
result = thaieda.run(df, llm=True, privacy="insight_only", provider="ollama")
print(result.llm_response)

# Or use OpenAI/Anthropic — still safe with insight_only
result = thaieda.run(df, llm=True, privacy="insight_only", provider="openai")

Privacy Modes

Control exactly what data leaves your machine:

Mode	What Leaves	Guarantee	When to Use
insight_only (default)	Stats + insights only	Raw data never leaves	Government, medical, PDPA data
anonymized	Data with PII → tokens	Names/phones/ID cards masked	Need structure without raw PII
dp_noise	Stats + Laplace noise	Prevents re-identification	Small datasets where stats leak
full	Everything	None — you accept the risk	Public data, demos

from thaieda.llm import analyze_with_llm

# Each mode as a standalone call
answer = analyze_with_llm(df, privacy="insight_only", provider="ollama")
answer = analyze_with_llm(df, privacy="anonymized", provider="openai")
answer = analyze_with_llm(df, privacy="dp_noise", provider="anthropic", epsilon=0.5)

Manual Pipeline (Full Control)

from thaieda import profile, discover_insights
from thaieda.detect import detect_all

# Step-by-step if you want control
report = profile(df, clean=True)
report.to_html("report.html")

result = discover_insights(df, detect_all(df), top_n=8)
for card in result.cards:
    print(f"[{card.pattern}] {card.description_th}")
    print(f"  → {card.recommendation_th}")

---

What ThaiEDA Catches

Problem	Example	What Happens
Buddhist Era dates	15/03/2567	Auto-detects พ.ศ. → converts to CE
Thai numerals	๑๒๓ in numeric column	Converts to 123
Zero-width spaces	สม\u200bชาย	Strips invisible chars
Mojibake encoding	à ¬Â¸Â¡Â¹	Auto-detects TIS-620 → UTF-8
Thai month names	มกราคม	Parses to ISO date
Phone numbers	081-234-5678	Detects + normalizes
National ID cards	1-1234-56789-01-2	Detects via regex
Placeholder values	-, N/A, ไม่มี	Flags as missing
Constant columns	All same value	Flags as useless

---

Installation

# Basic — works immediately
pip install thaieda

# Everything in one command
pip install "thaieda[all]"

# Or pick what you need
pip install "thaieda[thai]"        # Thai tokenizer
pip install "thaieda[ner]"         # Thai NER
pip install "thaieda[ml]"          # ML anomaly detection
pip install "thaieda[timeseries]"  # STL decomposition
pip install "thaieda[excel]"       # Excel support
pip install "thaieda[stats]"       # p-values (scipy)

# LLM providers (all optional, lazy-imported)
pip install openai                 # OpenAI GPT
pip install anthropic              # Anthropic Claude
pip install ollama                 # Ollama local LLM

Requirements: Python 3.10+, pandas, numpy, matplotlib, Jinja2

---

Modules

Module	What It Does
run() / EDA()	One-liner API — full pipeline in one call
io/	Auto-read CSV/JSON/JSONL/Excel + encoding detection
detect/	Column type detection + Thai month names
clean/	Encoding fix, numerals, BE→CE, dates, duplicates, missing
quality/	Thai quality checks + placeholder/constant detection
anomaly/	Statistical + ML + text anomaly detection
ner/	Thai NER: person/place/organization
insight_engine/	6 cross-column insight patterns (BH-corrected)
viz/	Auto charts with Thai font support
report/	Self-contained HTML report (Jinja2)
llm/	Privacy-preserving LLM analysis (4 modes, 3 providers)
schema/	Multi-file PK/FK discovery + relationship matching
timeseries/	Trend/seasonality/STL/ACF/gap detection

---

Testing

pytest tests/ -v              # all tests (424 passed)
pytest tests/test_oneliner.py # one-liner API tests
pytest tests/test_llm.py      # LLM + privacy mode tests
ruff check src/ tests/        # lint
ruff format src/ tests/       # format

---

License

Apache-2.0 © Peet Wannasarnmetha