Analytics for ML Features Dashboard

An interactive data science dashboard designed to analyze user behavior, model performance, and spending efficiency for a platform offering Machine Learning features. This tool provides actionable insights into how different license tiers interact with various ML models. You can visit the app at this link: https://analytics-for-ml.streamlit.app

---

📖 Project Overview

This project processes a complex dataset containing user activity logs, license types, and credit consumption. It transforms raw data into a visual story across five key dimensions:

• Usage Metrics: Tracking 1,866 unique users across various features.
• Cost Efficiency: Analyzing the spent_amount relative to requests_cnt.
• License Segmentation: Differentiating behaviors between Basic, Standard, Premium, and Enterprise tiers.
• Temporal Trends: Monitoring daily, weekly, and monthly growth.

---

Key Modules

1. Overview & Dataset Exploration

Initial high-level summary of the ecosystem.

• Metrics: Quick view of Total Users, Average Requests, and Total Spent.
• Distribution: Histograms showing Feature and License frequency.
• Heatmap: Correlation between specific Features and ML Models.

2. Relation Exploration

Deep dive into the relationship between license tiers and model usage.

• Box Plots: Visualizing request count and unit spending variance per model.
• Regression Analysis: Scatter plots with OLS trendlines showing the linear relationship between activity and costs.

3. Trends Over Time

Analyzing how user engagement evolves.

• Daily Correlation: Monitoring the stability of the usage-based pricing model.
• Weekly/Monthly Aggregates: Identifying "Working Day" peaks vs. "Weekend" troughs.
• Behavior Segments: Area charts showing the shift in user types (e.g., Core Users vs. Free Users).

4. User Behaviour Analysis

Focusing on retention and engagement funnels.

• Engagement Funnel: Tracking users from initial app use to high-credit spending.
• Retention: Histogram of "Days Active" grouped by license type.
• Power Users: Identifying the top 5 contributors to platform revenue per license.

5. Strategic Summary & KPIs

A final synthesis of all data points.

• KPI Expanders: Organized view of Efficiency, Engagement, and Power Usage metrics.
• Recommendations: Concrete business advice based on data patterns.

---

Key Business Insights

• Model Dominance: Models C and D are the platform's "workhorses," handling the majority of high-demand tasks (Feature 1 & 4).
• Usage-Based Pricing: A strong positive correlation ($r \approx 0.94$) proves that the spending model scales predictably with activity.
• The "Power User" Effect: A small minority of Premium users accounts for a disproportionate share of total spending.
• Retention Link: Higher-tier licenses (Premium/Enterprise) show a direct correlation with longer user lifespans.

---

Tech Stack

• Language: Python 3.x
• Data Handling: pandas, numpy
• Visualization: plotly.express, matplotlib
• Web Framework: streamlit
• UI Components: streamlit_option_menu, streamlit_lottie

---

Installation & Running

1. Clone the repository:

   git clone [https://github.com/your-username/ml-features-analytics.git](https://github.com/your-username/ml-features-analytics.git)
   cd ml-features-analytics