AI Data Analyst Agent

Pipeline Launcher

AI Data Analyst

Upload a dataset and let the multi-agent pipeline profile, analyze, validate, and explain it.

Backend connected

Drop your CSV here

or browse from your computer

CSV files up to 50MB

Live Pipeline Execution

0 of 5 stages complete

0%

  1. 1. Dataset Profiling

    Sweetviz
    —

    Scanning schema, distributions, and missing-value patterns.

    Status: pending

  2. 2. Plan Generation

    Planner Node
    —

    Drafting an analysis plan grounded in the dataset profile.

    Status: pending

  3. 3. Python Code Execution

    Sandbox
    —

    Running generated pandas / matplotlib analysis in an isolated sandbox.

    Status: pending

  4. 4. Insight & Recommendation Synthesis

    Executive Insight Agent
    —

    Converting verified evidence into executive narrative and actions.

    Status: pending

  5. 5. Report Compilation

    HTML & PDF
    —

    Assembling the deliverable report with charts and citations.

    Status: pending
Pipeline Configuration
Tune retry behaviour, creativity, and execution limits.
Max Retries3

Reflector attempts before a node is marked failed.

Temperature0.10

Lower values keep analysis deterministic and grounded.

Execution Timeout60s

Hard limit for sandboxed Python execution per node.

Sample datasets

Start instantly with a curated dataset — one click runs the full agent pipeline.

Sales Transactions

~30 rows · 10 columns

Regional transaction and revenue analysis.

Hiring Bias Data

~10 rows · 9 columns

Candidate demographics and hiring outcome analysis.

E-commerce Churn

~10 rows · 9 columns

Customer behavior and churn prediction analysis.