Zidio Internship Project v1.0

AI-Powered
Retail Intelligence

End-to-end analytics platform combining Prophet forecasting, LSTM deep learning, and XGBoost churn prediction — deployed with Docker, Kubernetes, and CI/CD.

6+ ML Models
1M+ Transactions
5 Dashboard Pages
28 Days of Dev

Revenue Analytics

Monitor daily revenue fluctuations and track the performance of your primary conversion funnels.

Demand Forecasting

+34%

Hybrid Prophet + LSTM ensemble model

Churn Prediction

Risk
0.94 ROC AUC
Accuracy 91.2%
Precision 89.4%
Recall 87.8%
XGBoost Optuna

Customer Segmentation

K-Means DBSCAN RFM

Inventory Optimization

Active
Stockout Prevention 94.7%
SKU Coverage 100%
Safety Stock Reorder Point

MLOps Pipeline

Operational

Full experiment tracking with automated retraining and drift detection

📥
Ingest
🧹
Clean
🤖
Train
📊
Evaluate
🔍
Drift
🚀
Deploy
MLflow Evidently Docker CI/CD

Upgrade to Pro

Deployed with Docker containers, Kubernetes orchestration, GitHub Actions CI/CD, and Prometheus + Grafana monitoring. Access the full live dashboard.

Launch Dashboard
🐳
Docker Containerized
☸️
Kubernetes Orchestrated
🔄
GitHub Actions CI/CD Pipeline
📡
Prometheus Monitoring

Active Campaigns

Production-grade tools powering the analytics pipeline

🐍 Python 3.11 Core
🔥 PyTorch Deep Learning
📊 Streamlit Dashboard
📈 Plotly Visualization
🤖 Scikit-Learn ML Framework
🔮 Prophet Forecasting
XGBoost Classification
🧪 MLflow Experiment Tracking
🐳 Docker Container
☸️ Kubernetes Orchestration
🔄 GitHub Actions CI/CD
📡 Prometheus Monitoring

Q3 Enterprise Push

Validated accuracy across all analytics modules

Demand Forecasting

MAPE ↓
ModelMAPEStatus
Prophet 28.3%
LSTM 26.1%
Hybrid Ensemble 24.1%

Churn Prediction

XGBoost
MetricValueStatus
Accuracy 91.2%
Precision 89.4%
Recall 87.8%
ROC AUC 0.94