📁 BACKSURE : Intelligent Backup Success Prediction System
"When data is everything, predicting its safety becomes the smartest move."
🔹INTRODUCTION
-During my 6th semester, I developed a deep curiosity about how data systems handle and protect massive information. While studying Machine Learning, especially algorithms like Logistic Regression, I often wondered, can data help us predict what might happen before it actually does?
-That curiosity led to BackSure: A machine learning–driven solution that predicts backup job success or failure using past backup logs.
-In a world where data loss can cost millions, predictive insights can save companies from failed backup operations and potential downtime.
🔹 WHY I BUILT THIS PROJECT?
-As a learner, I always loved exploring how systems think and make decisions.
-While performing machine learning lab experiments during my 6th semester, I realized that backup reliability is a real pain point in industries storage errors, network issues.
-So, BackSure became my way to connect:
-My machine learning learnings (from lab experiments), and
-My interest in system reliability & data management
-Into one practical, smart project.
🔹WHAT PROBLEMS DOES BACKSURE SOLVE??
▪️Backup Uncertainty: Predicts whether a backup will succeed or fail.
▪️ Data-Driven Decision Making: Uses backup history to learn and adapt continuously.Learns from past backup logs to improve future reliability.
▪️ Performance Monitoring: Tracks metrics like size, duration, and network latency to reveal performance bottlenecks.
▪️ Resource Optimization: Helps plan storage, time, and bandwidth more efficiently.
▪️ Informed Monitoring: Turns raw backup data into clear, actionable insights.
▪️ Scalable Insight: Works for small setups to enterprise-scale systems using any backup log CSV.
🔹 FEATURES
-Predictive Analysis – Uses ML to predict backup success or failure.
-Smart Visualization – Displays graphs & insights based on data trends.
-Logistic Regression Model – Learns from past backup patterns.
-Real-time Uploads – Upload a .csv backup log and get instant prediction.
🔹 TECH STACK
-Frontend = React + TypeScript + TailwindCSS + Framer Motion.
-Machine Learning= Python (Logistic Regression, Pandas, NumPy, Scikit-learn).
-Backend =Node.js + Express.js (handles ML prediction API).
-Data Format =CSV backup logs
🔹 WORKING
-User Uploads CSV File → Contains past backup data (features like size, duration, latency, etc.)
-Backend Sends Data to ML Model → Express.js forwards it to the trained Python model.
-Logistic Regression Predicts Success/Failure → Based on learned patterns (from historical backups).
-Frontend Displays Graph & Result Explanation → Green for successful backups, Red for likely failures.
🔹 HOW TO RUN IT LOCALLY
▪️ Clone Repository
-git clone https://github.com/Shriya-23/Backsure.git
-cd Backsure
▪️ Run Backend
-cd backend
-npm install
-node server.js
▪️ Run Frontend
-cd frontend
-npm install
-npm run dev
▪️ Run ML Model
-cd ml-model
-python analyze_backup.py ./data/backup_data.csv
📊 EXAMPLE CSV COLUMNS (Used for Prediction)
Here’s how BackSure reads and learns from your backup log data:
-job_id: Unique identifier for each backup job.
-size_gb: Total size of the backup file (in gigabytes).
-duration_min: Time taken to complete the backup (in minutes).
-error_count: Number of errors encountered during the process.
-retry_count: How many times the backup was retried.
-network_latency_ms: Network delay experienced during the backup (in milliseconds).
success: Target column — 1 means backup succeeded, 0 means it failed.
✨ CONCLUSION
BackSure combines the power of Machine Learning and System Reliability Engineering to make backups smarter, safer, and more predictable
As a learner, I built BackSure to explore how data-driven approaches can solve real operational challenges. This project taught me the importance of model interpretation, and system resilience.
💡 I’m always open to improving it further
Got suggestions or feedback? Reach out at [email protected]
— I’d love to learn from you!
💼 ABOUT ME
Hello! I’m Shriya Sharma,A Computer Science student passionate about building practical, data-driven, and impactful tech solutions.
I love transforming ideas into simple, meaningful tools that bridge the gap between technology and real-world problems.