Integrate tools such as DeepSeek into your daily analysis workflow, focusing on prompt creation, data analysis tasks, and result validation; ideal for operations, business, and data professionals who want to learn by doing.
Simply handing a spreadsheet to AI and receiving a professional-sounding explanation does not mean the data analysis is complete. What truly matters is whether the problem has been clearly defined, whether the data is consistent, whether the calculations are correct, and whether the conclusions can be substantiated by the evidence. *Data Analysis with AI* is ideal for readers who view AI as a work assistant but wish to retain their own judgment and verification capabilities.
Author: Xu Xiaolei; Tsinghua University Press; Published in August 2025; ISBN 9787302697756. The full title of the book is *Data Analysis with AI: Boosting Efficiency by a Factor of N Using DeepSeek, Doubao, and Zhipu Qingyan*. The reference to “efficiency” in the title is taken directly from the original book and does not constitute a guarantee by this website regarding the actual degree of efficiency improvement.
According to the official table of contents and preface, this book focuses on prompt engineering, task scenarios and goal setting, multi-round validation, and data analysis practices. The published table of contents lists case studies such as descriptive analysis, e-commerce sales data, and book bestseller list data, and discusses the performance of various tools. Some of the promotional text on the publisher’s webpage is inconsistent with the book’s themes; this entry is primarily based on the book title, bibliographic information, table of contents, and preface, and does not treat any inconsistencies as course commitments.
For readers in a professional setting, what matters is not whether a particular prompt can be copied, but whether it clarifies the task context, field definitions, scope of analysis, output format, and validation methods. If these elements are unclear, switching to a different model may not necessarily yield reliable answers.
If the goal is to gain a deeper understanding of statistical inference, experimental design, or complex modeling, additional courses in these areas are required. Tools can assist in the process, but they cannot replace an understanding of metrics, samples, and the business context.
Select a set of practice data that does not contain personal information or trade secrets, and start by writing a field dictionary and three clear questions. Have the AI explain the analysis steps, then ask it to provide a reproducible formula or code. Next, use a spreadsheet or other reliable tool to verify the key figures, and finally categorize the conclusions into three groups: “directly supported by the data,” “requires additional evidence,” and “cannot be determined at this time.”
For example, an increase in sales may be due to changes in average order value, order volume, or the scope of the analysis. As you read, keep asking yourself: Has the denominator changed? How were missing values handled? Has a correlation been mistaken for a causal relationship? Only by clearly addressing these questions can a report effectively support decision-making.
Do not copy model-generated numbers directly into formal reports; do not overlook data definitions just because the charts look good; and do not upload customer information, employee records, or internal operational data without authorization. In real-world work scenarios, you should first confirm your organization’s data usage policies and, whenever possible, use anonymized samples for validation.
The comparisons of specific tools in this book reflect the environment at the time of writing. Model versions, account permissions, and file-handling capabilities are subject to change; the current user experience requires retesting, so a single comparison should not be regarded as a permanent ranking.
If you’re most concerned with the spreadsheets, reports, and business issues at hand, this book will be more directly applicable to your specific tasks; if you want to understand model training and deployment, you should prioritize *Large-Scale Language Models: From Theory to Practice (2nd Edition)*; if you lack a foundation in AI concepts, you may want to start with *The Underlying Logic of Artificial Intelligence*. The best way to approach this book is to work through the tasks hands-on and keep a record of your own verification results.
Date of bibliographic information verification: October 17, 2026. This site does not claim to provide current sales rankings, nor does it promise a fixed efficiency multiplier; the exercises and risk warnings listed above are editorial recommendations.
The cover is used solely for the purpose of identifying and introducing this book; copyright belongs to the respective rights holders.
简体中文 简体中文 
English