InsightBot

Automated data insight generator — analyzes CSV/JSON datasets and produces natural language insights about trends, outliers, and patterns.

Inspired by data analytics AI trends.

---

Architecture

graph LR
    A[CSV / JSON Data] -->|load| B[InsightBot Core]
    B --> C[Statistical Analysis]
    C --> D[Outlier Detection]
    C --> E[Trend Analysis]
    C --> F[Correlation Finder]
    D --> G[Natural Language Engine]
    E --> G
    F --> G
    G --> H[Insight Report]
    H --> I[Text / JSON / Markdown]

Loading

Quickstart

Install

pip install -e .

Usage

from insightbot import InsightBot

bot = InsightBot()
bot.load_csv("sales.csv")

# Get a statistical summary of a column
summary = bot.describe("revenue")
print(summary)

# Detect outliers using IQR method
outliers = bot.detect_outliers("revenue")
print(f"Found {len(outliers)} outliers")

# Find correlations between numeric columns
correlations = bot.find_correlations()
for corr in correlations:
    print(corr)

# Find trends in a column
trends = bot.find_trends("revenue")
print(trends)

# Generate all insights at once
insights = bot.generate_insights()
for insight in insights:
    print(f"[{insight.category}] {insight.message}")

# Export a full report
bot.export_report(format="markdown", path="report.md")

CLI

# Quick summary
python -m insightbot data.csv

# Export JSON report
python -m insightbot data.csv --format json --output report.json

Features

• CSV & JSON ingestion — load tabular data from common formats
• Statistical summaries — mean, median, std dev, min/max, percentiles
• Outlier detection — IQR-based identification of anomalous values
• Trend analysis — detect increasing, decreasing, or stable trends
• Correlation discovery — find relationships between numeric columns
• Natural language insights — human-readable descriptions of findings
• Export formats — text, JSON, and Markdown reports

Development

make install    # install in dev mode
make test       # run tests
make lint       # run linter
make format     # format code

License

MIT — see LICENSE for details.

---

Built by Officethree Technologies | Made with ❤️ and AI