README.md

---
title: Transformer Oil Temperature Forecaster
emoji: ⚡
colorFrom: red
colorTo: blue
sdk: gradio
sdk_version: 6.12.0
app_file: app.py
pinned: false
---

# ⚡ Transformer Oil Temperature Forecaster

> **ARIMAX · Anomaly Detection · Time Series Analysis**

Upload ETT-style transformer CSV data and get:

| Feature | Details |
|---|---|
| **Model** | ARIMAX — auto-selects best `(p, d, q)` via AIC grid search |
| **Endog** | `OT` — oil temperature |
| **Exog** | `HUFL, HULL, MUFL, MULL, LUFL, LULL` — load features |
| **Stationarity** | ADF test; auto-applies 1st differencing if needed |
| **Anomaly Detection** | Residual-based, threshold = mean ± 2.5σ |
| **Evaluation** | MAE + RMSE on 20% hold-out set |

---

## 📂 Expected CSV Format

```
date,HUFL,HULL,MUFL,MULL,LUFL,LULL,OT
2016-07-01 00:00:00,5.827,2.009,1.599,0.462,4.203,1.340,30.531
...
```

The ETT (Electricity Transformer Temperature) dataset works out of the box.  
Download it from: https://github.com/zhouhaoyi/ETDataset

---

## 🚀 Running Locally

```bash
pip install -r requirements.txt
python app.py
```

---

## 📐 Architecture

```
CSV Upload
    │
    ▼
load_data()          ← parse datetime index, ffill missing
    │
    ▼
check_stationarity() ← ADF test → d value
    │
    ▼
train_arimax()       ← grid search (p,q) on 80% train split
    │
    ├──► forecast()          ← out-of-sample N steps
    │
    └──► detect_anomalies()  ← residual threshold flagging
```