Data Analysis

SaaS, open source and skills for data analysis work — with the limits of each written down.

Before you pick

What decides it

Whether your definitions live outside the chat decides this. Otherwise each question re-infers which column means revenue and which join is right; answers drift.

What to avoid

Wrong answers don't error: the query runs, the chart renders, the number is plausible. Verifying every result by hand costs back the time saved.

What is changing

Write a query, run it, read what came back, rewrite: accuracy now comes from that loop, and single-shot generation no longer competes.

SaaS

Hosted products you sign up for.

Leading

Established picks, at the top of today's aggregate.

1ChatGPT (Advanced Data Analysis)The default for ad-hoc analysis on an uploaded file — writes and runs Python, charts, and explains results conversationally.
2Julius AIConversational analysis of spreadsheets and CSVs with expert-level insights and visualizations in seconds.
3HexAI-assisted notebooks for data teams that live in SQL and Python and want reproducible, shareable analysis.
4Tableau (with Einstein/Pulse AI)Enterprise BI dashboards with natural-language querying when your data already lives in a governed warehouse.
5Microsoft Excel + CopilotEnterprise AI inside Excel itself — formulas, pivots, and trend explanations in plain English without leaving the sheet.

Emerging

Newer challengers, surfaced by two or more sources.

1QuantumLayers
2Anomaly AIConnects spreadsheets, ad platforms and warehouses, then cleans the data and builds refreshable dashboards for reports you rerun every month.

Open source

Repositories you run yourself.

Leading

1pandasThe de-facto foundation of data analysis in Python (49k★) — the DataFrame library that ChatGPT Advanced Data Analysis and nearly every AI analysis tool run under the hood.
2Apache SupersetThe most-starred open-source BI/exploration platform (74k★) — a self-hostable Tableau/Looker alternative for SQL-driven dashboards and ad-hoc data exploration.
3MetabaseThe easiest open-source BI tool to self-host (48k★) — no-SQL question builder plus full SQL for teams that want fast dashboards over their database.
4Project Jupyter (JupyterLab)The universal notebook environment where interactive analysis actually happens (15k★, ACM Software System Award project) — Python/R/Julia cells, charts, and reproducible reports.
5PostHogOpen-source product analytics (37k★) — event data, funnels, session replay, and SQL when your analysis is about how users behave in a product.

Emerging

1georgekhananaev/excel-ai-assistantDesktop app that applies AI prompts to selected Excel or CSV cells for cleaning and reformatting, via OpenAI or local Ollama models.
2shubham303/meelu-analytics-mcp

Skills

Skills you drop into a coding agent.

Leading

1xlsx (Anthropic Agent Skill)Anthropic's official spreadsheet skill (from the 163k-star anthropics/skills repo, used in production Claude) — reads, analyzes, and builds Excel with formulas, pivots, and charts; the recognized leader for AI file-based analysis.
2K-Dense scientific-agent-skills (EDA)Automated exploratory-data-analysis and publication-quality visualization skills from the most-starred science skills library (31.6k★, 148 SKILL.md skills).
3nimrodfisher/data-analytics-skillsA small, focused community collection of Claude skills for the everyday analytics workflow — cleaning, aggregation, and charting on tabular data. Low star count, so ranked low.
4juanlurg/data-science-claude-skillsEnvironment-aware EDA skills (auto-eda, dataset-doctor) that detect your pandas/polars library and SQL dialect. Genuine but niche and low-star, so ranked low.

Emerging

Nothing here we could place with confidence.