📈 AI Insights & Analytics

Sentio's AI can automatically generate metrics queries, create visualizations, and provide intelligent analysis of your blockchain data patterns and trends.

Overview

The Insights & Analytics AI goes beyond simple SQL generation to provide:

Metrics-First Approach: Focus on KPIs and business metrics
Smart Visualizations: Automatic chart type selection and configuration
Time-Series Analysis: Advanced temporal pattern recognition
Multi-Dimensional Analysis: Compare metrics across different dimensions
Anomaly Detection: Identify unusual patterns and outliers

Core Capabilities

Metrics Generation

Natural Language: "Track daily active users for my DeFi protocol"

Generated Insight:

Metric: Daily Active Users
Query: COUNT(DISTINCT user_address)
Time Grain: Daily
Filters: Successful transactions only
Chart: Line chart with trend analysis

Time-Series Analysis

Natural Language: "Show me TVL trends with weekly breakdown"

Generated Analysis:

Base Metric: Total Value Locked
Aggregation: Weekly average
Time Range: Last 3 months
Trend Analysis: 15% growth rate
Seasonality: Weekend dips identified
Chart: Area chart with moving average

Comparative Analysis

Natural Language: "Compare trading volumes between different token pairs"

Generated Comparison:

Primary Metric: Trading Volume (USD)
Dimensions: Token Pair, Time Period
Breakdowns: Top 10 pairs by volume
Analysis: USDC/ETH dominates with 35% share
Chart: Stacked bar chart with percentage view

Supported Metric Types

Volume Metrics

Trading Volume, Transaction Volume, Transfer Volume, Liquidity Volume

Activity Metrics

Active Users (DAU/WAU/MAU), Transaction Count, Contract Interactions, Sessions

Financial Metrics

Revenue, TVL, Market Cap, Yield

Performance Metrics

Success Rate, Averages (gas, size), Latency, Efficiency

Advanced Analytics Features

Cohort Analysis

Natural Language: "Show me user retention cohorts for new wallet addresses"

Generated Cohort Analysis:

Cohort: Users by first transaction month
Retention: Return within 30/60/90 days
Visualization: Cohort retention heatmap
Key Insight: 60% of users return within 30 days

Funnel Analysis

Natural Language: "Create a funnel from wallet connection to first trade"

Generated Funnel:

Wallet Connected → Token Approved → Trade Initiated → Trade Completed
Conversion Rate: 42% overall

Segmentation Analysis

Natural Language: "Segment users by trading frequency and volume"

Generated Segments:

Whales, Active Traders, Casual Users, Dormant

Anomaly Detection

Natural Language: "Find unusual spikes in gas usage"

Detected Anomaly Example:

Date: 2024-03-15 14:00 UTC
Metric: Average Gas Usage
Observed: 180,000 (300% above normal)
Likely Cause: Popular NFT mint event

Visualization Intelligence

Automatic chart selection based on data
Optimized configuration (colors, axes, annotations, tooltips)
Supports time series, categorical, distribution, and correlation views

Multi-Dimensional Analysis

Drill-down from protocol → pool → token → user levels
Cross-metric correlation (e.g., gas price vs transaction volume)
Market context integration (price data, events, network status)

Best Practices

Start with business questions and success metrics
Provide relevant context (mechanics, changes, market conditions)
Validate results against external sources and historical patterns
Iterate and refine insights over time

Dashboard Integration

Automated dashboard creation (executive, operational, deep dive, alerts)
Real-time monitoring with threshold, trend, anomaly, and competitive alerts
Scheduled reports (daily/weekly/monthly/quarterly)

Troubleshooting

"Not enough data points": Extend time range or adjust granularity
"Metrics look inconsistent": Verify methodology and data quality
"Charts are too cluttered": Limit categories/series or split charts