Inquira 🔍

Inquira is an AI-powered product sales predictor built with Flask. It analyzes product pricing, discount metrics, review sentiment, and product descriptions to predict whether a product will have Good Sales or Low Selling performance — and if underperforming, explains why using SHAP and gives actionable suggestions to improve.

---

Features

• Sales Prediction — Classifies a product as "Good Sales" or "Low Selling" using a trained ML pipeline.
• SHAP Explainability — For low-selling predictions, surfaces the top contributing factors driving the result.
• Actionable Suggestions — Provides concrete recommendations to improve product performance.
• NLP Description Analysis — Analyzes product descriptions using TextBlob and Gensim to identify language quality issues.
• Flask Web Interface — Simple, clean web UI for entering product details and viewing results.

---

Project Structure

Inquira/
├── app.py                   # Flask app — routes and request handling
├── predictor.py             # Loads pipeline.pkl and runs predictions
├── explain.py               # SHAP-based explanation and suggestions
├── description_analyzer.py  # NLP analysis of product descriptions
├── train_pipeline.py        # Script to train and save the ML pipeline
├── pipeline.pkl             # Trained scikit-learn pipeline (serialized)
├── amazon_clean_data.csv    # Cleaned Amazon product dataset
├── sales.ipynb              # Exploratory data analysis notebook
├── style.css                # Stylesheet
├── templates/               # Jinja2 HTML templates
└── requirements.txt         # Python dependencies

---

Input Features

Feature	Description
discounted_price	Current selling price after discount
actual_price	Original MRP of the product
discount_percentage	Percentage discount applied
rating	Average customer rating
discount	Absolute discount amount
price_diff	Difference between actual and discounted price
description	Product description text (auto-computes length)
review_sentiment	Sentiment score derived from customer reviews

---

Getting Started

1. Clone the repository

git clone https://github.com/Prriiiyankaaa/Inquira.git
cd Inquira

2. Install dependencies

pip install -r requirements.txt

3. (Optional) Retrain the model

python train_pipeline.py

Skip this step if you want to use the pre-trained pipeline.pkl included in the repo.

4. Run the app

python app.py

Then open your browser and navigate to http://127.0.0.1:5000.

---

Tech Stack

Layer	Tools
Web Framework	Flask
ML	scikit-learn, SHAP, joblib
NLP	TextBlob, Gensim
Data	pandas, numpy
Visualization	matplotlib, seaborn
Dataset	Amazon product data (amazon_clean_data.csv)

---

How It Works

1. The user submits product details via the web form.
2. predictor.py loads the serialized pipeline.pkl and returns a prediction.
3. If the product is predicted as Low Selling:
   • explain.py uses SHAP to identify which features are hurting performance and generates suggestions.
   • description_analyzer.py runs NLP analysis on the product description to flag quality issues.
4. Results — prediction, SHAP reasons, suggestions, and description analysis — are rendered in the UI.

---

Dataset

The project uses amazon_clean_data.csv, a cleaned dataset of Amazon product listings with fields including pricing, ratings, discounts, and review data. The notebook sales.ipynb contains the full EDA and feature engineering process.

---

Contributing

Pull requests are welcome! For major changes, please open an issue first to discuss what you'd like to change.

---

Author

Prriiiyankaaa — GitHub