Shaolei Zhang

Assistant Professor at Renmin University of China
Prev: Ph.D. from Institute of Computing Technology, Chinese Academy of Sciences

LLM | Multimodal LLM | Agent

[sitin @sitinme]

做数据分析的人应该都懂那种崩溃感：
Excel 一开就卡、SQL 跑半天、模型调参像瞎猜、报告一写就是一下午。
更别说很多人本职工作也不是“数据科学家”，但又不得不啃一堆 CSV、报表和图表。

这两天看到人民大学 RUC Datalab 开源的DeepAnalyze，有点眼前一亮。

它不是那种“帮你写点分析文案”的 AI，而是想做一个真正的「AI 数据分析师」：

能读 Excel/CSV/JSON/TXT，自己规划步骤、清洗数据、建模、画图、最后给你一份完整报告。

甚至还能把推理链（reasoning trace）展示出来，让你看到它是怎么一步步想出来的。

我自己最有感觉的点有几个：
以前 ChatGPT 更像“写稿子、写代码”的助手，但对数据本身不够“懂”；
DeepAnalyze 直接把这一块补上了，做到“从数据到报告”的全流程；
对普通人来说，可能第一次有机会把“专业的数据分析能力”当成一个工具来用，而不是一门必须自己啃几年的技能。

当然，它现在肯定还不完美，落地也需要折腾环境、准备数据。

但这个方向我很看好：
AI 帮你做繁琐的统计和建模，人类负责问好问题、选好指标、做最后决策。
数据分析不再是少数人的特权，而是变成人人可用的一种“智能能力”。

简单说：
过去是「你学会数据分析再去用工具」，
以后可能是「你先有一个 AI 数据分析师，再慢慢学会怎么问对问题」。

[Shaolei Zhang @Vily_1998]

@sitinme 感谢分享~欢迎大家试用DeepAnalyze，目前已可以直接在Linux/Mac通过WebUI和JupyterUI使用。上传文件+一句指令，直接生成数据分析报告，作为您的专属数据分析助手😊。
论文：https://t.co/zgSFzhPMpH
代码、Demo：https://t.co/1SeX5mulRh

[机器之心 JIQIZHIXIN @jiqizhixin]

AI can become a fully autonomous data scientist.

DeepAnalyze-8B is the first agentic LLM capable of handling the entire data science pipeline—from raw data to analyst-grade research reports—without predefined workflows. It learns like a human via a curriculum-based agentic training paradigm and a data-grounded trajectory synthesis process.

Despite having just 8B parameters, DeepAnalyze surpasses workflow-based agents built on proprietary LLMs, marking a major step toward open, autonomous data science.

DeepAnalyze: Agentic Large Language Models for Autonomous Data Science

RUC, Tsinghua
Paper: https://t.co/NGAqY6Bp1l
Code: https://t.co/oqvDYonyDU

Our report: https://t.co/p0Icb7TiJD

📬 #PapersAccepted by Jiqizhixin

[Shaolei Zhang @Vily_1998]

@jiqizhixin Thanks to Jiqizhixin for the share! DeepAnalyze has gained over 1,600 GitHub ⭐️ in just two weeks — give it a try!

[Shaolei Zhang @Vily_1998]

@rohanpaul_ai DeepAnalyze has 1300+ GitHub⭐️. The model, code, and demo of DeepAnalyze are all open-sourced.

DeepAnalyze can be deployed locally as your personal AI data analyst 🤖, exploring data freely and generating analyst-grade research reports 📈.

Code&Demo: https://t.co/1SeX5mulRh

[Shaolei Zhang @Vily_1998]

DeepAnalyze开源一周，GitHub收获1.2K⭐！
DeepAnalyze是首个面向数据科学的 Agentic LLM，自主完成任意数据科学任务:
🛠 数据任务：数据准备、分析、建模、可视化及洞察
🔍 数据研究：在各种数据上深度研究，生成分析师级报告
论文: https://t.co/zgSFzhPMpH
代码、Demo: https://t.co/1SeX5mulRh

[Python Hub @PythonHub]

DeepAnalyze: Agentic Large Language Models for Autonomous Data Science

DeepAnalyze is the first agentic LLM for autonomous data science, supporting:
🛠 Data preparation, analysis, modeling, visualization, and insight.
🔍 Data research and produce research reports.

https://t.co/E9AtQSz3ma

[Shaolei Zhang @Vily_1998]

"Data Agent": emerging paradigm or overstated hype?

We introduce the first L0–L5 autonomy taxonomy—defining what data agents really are.

📈 Hugging Face Daily Papers Top 3.

🔗 Repo: https://t.co/eB1cIjg6rj
📚 Paper: https://t.co/XT4ZRbeRmY

#DataAgents #LLM #DataIntelligence

[Shaolei Zhang @Vily_1998]

@singularity20xy Thanks a lot for sharing our work!
Welcome to try DeepAnalyze at https://t.co/1SeX5mulRh, and we are happy to respond to any questions😊.

[Shaolei Zhang @Vily_1998]

@FMackenzie7 Thanks a lot for sharing our work!
Welcome to try DeepAnalyze at https://t.co/1SeX5mulRh, and we are happy to respond to any questions😊.

[Shaolei Zhang @Vily_1998]

@javaeeeee1 Thanks a lot for sharing our work!
Welcome to try DeepAnalyze at https://t.co/1SeX5mulRh, and we are happy to respond to any questions😊.

[Shaolei Zhang @Vily_1998]

@GitHub_Daily DeepAnalyze是首个面向数据科学的agentic大模型，完全开源、欢迎体验！同时欢迎大家在GitHub Repo讨论和贡献~
论文：https://t.co/zgSFzhPMpH
代码、Demo：https://t.co/1SeX5mulRh

[GitHubDaily @GitHub_Daily]

平时分析数据，从清洗、探索到建模、可视化，再到撰写报告，整个流程非常耗时，当要面对多个不同格式的数据文件时更是头疼。

恰巧，找到了 DeepAnalyze 这个开源项目，它能让 AI 自主完成整个数据科学流程，真正做到无需人工干预。

基于 DeepSeek-R1 和课程学习训练方法开发，支持数据准备、分析、建模、可视化到报告生成的全流程。

可以处理数据库、CSV、Excel、JSON、XML 等各种格式数据，最终直接生成专业级的研究报告给我们。

GitHub：https://t.co/fziwEeT1GT

同时提供 50 万条训练数据集，完全开源模型和代码，方便我们二次开发。

只需克隆代码到本地并下载模型后，通过 vLLM 部署即可使用。

[Shaolei Zhang @Vily_1998]

@AdinaYakup DeepAnalyze supports:
🛠 Data preparation, analysis, modeling, visualization, and insight.
🔍 Open-ended data research, produce analyst-grade research reports.

Welcome to try DeepAnalyze at our GitHub repo https://t.co/1SeX5mulRh, and any discussion or feedback is appreciated!

[Shaolei Zhang @Vily_1998]

DeepAnalyze supports:
🛠 Data preparation, analysis, modeling, visualization, and report generation.
🔍 Open-ended data research and produce analyst-grade research reports.

Welcome to try DeepAnalyze at our GitHub repo, and any discussion or feedback is appreciated!

[Adina Yakup @AdinaYakup]

DeepAnalyze🔥Agentic LLM for autonomous data science from RUC data Lab.

Model: https://t.co/0ldZBmJIxs
Dataset: https://t.co/dCUyv7G52e
Paper: https://t.co/dlqciWTfMi

✨ Fully open-source
✨ End-to-end automation
✨ Handles all data types

[Adina Yakup @AdinaYakup]

DeepAnalyze🔥Agentic LLM for autonomous data science from RUC data Lab.

Model: https://t.co/0ldZBmJIxs
Dataset: https://t.co/dCUyv7G52e
Paper: https://t.co/dlqciWTfMi

✨ Fully open-source
✨ End-to-end automation
✨ Handles all data types

[Shaolei Zhang @Vily_1998]

@aigclink Thanks a lot for sharing our work!
Welcome to try DeepAnalyze at https://t.co/1SeX5mulRh, and we are happy to respond to any questions😊.

[Shaolei Zhang @Vily_1998]

@Dr_Singularity Thanks a lot for sharing our work!
Welcome to try DeepAnalyze at https://t.co/1SeX5mulRh, and we are happy to respond to any questions😊.

[Shaolei Zhang @Vily_1998]

@rohanpaul_ai Thanks a lot for sharing our work!
Welcome to try DeepAnalyze at https://t.co/1SeX5mulRh, and we are happy to respond to any questions😊.

[Gorden Sun @Gorden_Sun]

DeepAnalyze：数据分析智能体
能自主完成一系列数据任务，包括：准备、分析、建模、可视化、生成报告。
Github：https://t.co/MujCqLvsKR