data2paper 是什么服务？

data2paper 是一项科研支持服务。您提供研究问题和数据说明，我们提供统计分析、图表生成、论文初稿支持、参考文献核验与投稿材料整理，交付内容需由研究者审核修改后使用。

数据安全如何保障？

评估阶段仅需元数据问卷，不需提交原始数据。正式合作后数据加密传输、本地分析，全程不上传任何第三方云服务。交付完成后7天内销毁数据。

服务价格是多少？

我们先提供免费评估，并在正式开始前基于研究类型、数据复杂度、交付范围和时间要求给出书面报价与里程碑。客户确认报价后才进入付费服务。

参考文献会不会是AI编造的？

绝不会。所有参考文献均附带 DOI 或 PubMed 链接，可逐条验证真实性。我们不编造数据、引用、伦理审批或研究结论。

退款和取消政策在哪里？

退款、争议、取消、数据处理、法律限制和促销条款列在本页 Policies 部分。

data2paper - AI辅助科研支持服务 | AI-Assisted Research Support

科研路上的技术阻碍

Overcoming Technical Barriers

统计不会做

Statistics Are Hard

SPSS、R、Python——方法选错了投稿直接被拒。不是不努力，是临床科研里的统计门槛实在太高。

SPSS, R, Python — choose the wrong method and your submission is rejected outright. The statistical bar in clinical research is simply too high.

"审稿人问我为什么用独立样本t检验而不是Mann-Whitney，我根本不知道怎么答。" "The reviewer asked why I used an independent t-test instead of Mann-Whitney. I had no idea how to respond."

论文写不出

Writer's Block

数据有了，结论也清楚，但Introduction怎么铺垫、Discussion怎么展开——每次对着空白页就卡住。

You have the data and the conclusions, but how to frame the Introduction, how to develop the Discussion — every time you face a blank page, you freeze.

"数据分析三天，写Introduction写了两周，还不知道哪里不对劲。" "Data analysis took three days. The Introduction took two weeks — and I still couldn't figure out what was wrong."

时间不够用

Never Enough Time

门诊、值班、基金申请、学生答辩……科研写作只能在周末深夜挤。晋升周期就在眼前，时间不等人。

Clinics, on-call shifts, grant applications, student defenses... Research writing is squeezed into late nights and weekends. Promotion deadlines wait for no one.

"副主任医师评审下个季度，手里三个在研项目，一篇SCI都没发出去。" "My associate chief physician review is next quarter. Three ongoing projects, zero SCI publications."

全方位的技术支撑

Comprehensive Technical Support

数据探查与清洗

Data Profiling & Cleaning

系统检测缺失值、异常值、分布特征，并生成数据质量报告。

Systematic detection of missing values, outliers, and distribution characteristics, with a data quality report.

统计分析

Statistical Analysis

方法选择有据可依，输出完整统计结果表格，附分析代码与方法说明。

Evidence-based method selection, complete result tables, with analysis code and methodology notes.

期刊级图表生成

Journal-Ready Figures

300 DPI图表输出，支持PNG/PDF/TIFF格式，便于研究者按目标期刊要求复核和调整。

300 DPI figure outputs in PNG/PDF/TIFF for researcher review and adjustment against target journal requirements.

论文初稿

Manuscript Draft

按IMRAD结构输出完整初稿，涵盖Introduction、Methods、Results、Discussion。

Complete IMRAD-structured draft covering Introduction, Methods, Results, and Discussion.

参考文献整理

Reference Management

全面检索相关文献，按目标期刊格式(AMA/Vancouver/APA等)输出引用列表。

Comprehensive literature search with citations formatted to target journal style (AMA/Vancouver/APA, etc.).

投稿材料准备

Submission Package

Cover Letter起草、Author Contribution声明、期刊格式核对，按目标刊要求整理。

Cover letter drafting, author contribution statements, and journal format checks tailored to your target venue.

真实案例：从 CSV 到可复核分析

Real Case: From CSV to Reviewable Analysis

公开数据集 · CC BY 4.0 · 299 例心衰患者 · 13 个临床变量 Open Dataset · CC BY 4.0 · 299 Heart Failure Patients · 13 Clinical Variables

步骤 1-2Step 1-2

数据探查 + Table 1 Data Profiling + Table 1

系统辅助识别分布类型，并由分析人员复核方法选择 System-assisted distribution checks with analyst review of method selection

步骤 3-4Step 3-4

生存分析 + Cox 回归 Survival Analysis + Cox Regression

KM 曲线 + Log-rank 检验 + 单因素筛选 → 多因素 Cox 回归森林图 KM curves + Log-rank + Univariate screening → Multivariate Cox forest plot

步骤 5-6Step 5-6

亚组分析 + 数字核对 Subgroup Analysis + Number Verification

6 维度亚组森林图 + 全文数字交叉核对，摘要=正文=表格=图 6-dimension subgroup forest plot + cross-document number reconciliation

按射血分数分层的 Kaplan-Meier 生存曲线 (Log-rank P=0.0022) KM Survival Curve by Ejection Fraction (Log-rank P=0.0022)

多因素 Cox 回归森林图 — 独立预测因子一目了然 Multivariate Cox Regression Forest Plot — Independent Predictors at a Glance

稿件片段预览 (Abstract) Manuscript Excerpt (Abstract)

Background: Heart failure remains a leading cause of mortality worldwide. Identifying independent predictors of survival may guide clinical decision-making. Methods: We retrospectively analyzed 299 heart failure patients. Kaplan-Meier curves with log-rank tests assessed survival differences across subgroups. Multivariate Cox proportional hazards regression identified independent predictors. Results: During a median follow-up of 115 days, 96 patients (32.1%) died. Age (HR=1.04, 95%CI 1.03-1.06), serum creatinine (HR=1.36, 95%CI 1.18-1.55), and hypertension (HR=1.61, 95%CI 1.06-2.44) were independent risk factors, while ejection fraction was protective (HR=0.95, 95%CI 0.94-0.97)...

背景：心力衰竭仍是全球主要死亡原因之一。识别独立预后预测因子有助于指导临床决策。 方法：回顾性分析 299 例心衰患者。采用 Kaplan-Meier 曲线及 Log-rank 检验评估亚组间生存差异，多因素 Cox 比例风险回归识别独立预测因子。 结果：中位随访 115 天，96 例（32.1%）死亡。年龄（HR=1.04, 95%CI 1.03-1.06）、血清肌酐（HR=1.36, 95%CI 1.18-1.55）和高血压（HR=1.61, 95%CI 1.06-2.44）为独立危险因素，射血分数为保护因素（HR=0.95, 95%CI 0.94-0.97）……

* 以上为示例稿件摘要片段，仅供展示。完整稿件需研究者审核、修改并确认后方可投稿。 * Example abstract excerpt for demonstration only. Full manuscript requires researcher review, revision, and approval before submission.

6
论文图表Manuscript Figures

5
统计表格Statistical Tables

1
可复现代码Reproducible Script

0
你需要写的代码行数Lines of Code You Write

下载完整 Demo 分析包 Download Full Demo Package

含：原始数据 + 分析代码 + 全部图表 + 统计表格 (CC BY 4.0) Includes: raw data + analysis script + all figures + statistical tables (CC BY 4.0)

眼见为实：每一步都经得起检验

See for Yourself: Every Step Stands Up to Scrutiny

以下展示均来自真实 Demo Case 的实际产出物，非模拟截图。 All screenshots below are actual outputs from our real demo case — not mockups.

数据安全：评估阶段不碰原始数据

Data Security: No Raw Data Needed for Assessment

全程本地运算，交付后 7 天内销毁 All computation runs locally. Data deleted within 7 days.

评估阶段仅需填写数据问卷（样本量、变量类型等元信息），不要求提交原始数据。正式合作如需处理原始数据，会先确认书面范围、脱敏要求、传输方式、访问权限和交付后处理规则。

During assessment, only a metadata questionnaire is needed — no raw data required. If raw data is required after engagement, we first confirm the written scope, de-identification requirements, transfer method, access permissions, and post-delivery handling terms.

统计方法透明：每步分析附选择理由

Transparent Methods: Every Step Includes Selection Rationale

系统辅助初筛，人工复核方法选择——不是黑箱 System-assisted screening with analyst-reviewed method selection — not a black box

交付物包含方法选择说明：变量分布、前提检验、备选方法和最终选择理由都会尽量写清楚，方便研究者和统计顾问复核。

Deliverables include method-selection notes: variable distributions, prerequisite checks, alternatives, and selection rationale are documented for researcher or statistician review.

全文数字一致性：摘要=正文=表格=图

Number Consistency: Abstract = Text = Table = Figure

全文数字交叉核对，附核对清单 Cross-document verification with reconciliation checklist

每份交付稿件都包含数字核对表，确保样本量、HR、P 值等关键数字在摘要、正文、表格和图表中完全一致。上图展示的就是 Demo Case 的真实核对结果——6 项核心指标，全部 PASS。

Every delivered manuscript includes a number reconciliation table ensuring that sample sizes, HRs, P-values, and other key metrics are identical across abstract, text, tables, and figures. The screenshot above shows real verification from our demo case — 6 core metrics, all PASS.

参考文献可验证：每条引用附 DOI 链接

Verifiable References: Every Citation Includes DOI Link

每条文献尽量附 DOI 或 PubMed 链接，便于逐条核验 Citations include DOI or PubMed links where available for one-by-one verification

参考文献需要能回到可核验来源。我们不接受编造引用；研究者在最终投稿前仍需逐条确认引用是否真实、相关且符合目标期刊格式。

References should trace back to verifiable sources. Fabricated citations are not accepted; researchers still need to confirm authenticity, relevance, and journal formatting before submission.

交付代码，可复现可修改

Full Code Delivery: Reproducible & Modifiable

交付完整分析代码，改参数重跑即可——不是黑箱 Complete analysis code delivered. Change parameters and re-run — not a black box.

交付可包含完整 Python 分析脚本，用于复现图表和统计结果。导师或统计顾问要求调整分组阈值时，可以在脚本中修改参数并重新运行，便于复核和迭代。

Delivery can include Python analysis scripts for reproducing figures and statistical results. When a supervisor or statistician asks for a different threshold, parameters can be adjusted and re-run for review and iteration.

亲自验证：下载完整 Demo 分析包

Verify Yourself: Download the Full Demo Package

包含公开示例数据 + 分析代码 + 图表 + 统计表格 + 数字核对表 (CC BY 4.0)

Includes public example data + analysis code + figures + statistical tables + reconciliation checklist (CC BY 4.0)

下载 Demo 分析包 (ZIP) Download Demo Package (ZIP)

我们的定位：data2paper帮你解决"统计方法不熟、写作耗时极长"这两个技术性障碍。研究设计、数据来源、结论解读与学术署名，始终是研究者的责任与权利。初稿是起点，不是终稿——所有统计结论与临床解读必须由研究者本人审核确认后方可投稿。我们是你的科研助手，不是你的共同作者。

Our Position: data2paper removes two technical barriers — unfamiliar statistics and time-consuming writing. Research design, data provenance, interpretation, and authorship remain the researcher's responsibility and right. The draft is a starting point — all conclusions must be reviewed by you before submission. We are your research assistant, not your co-author.

商家信息与服务政策

Business Information & Service Policies

以下信息用于说明 data2paper 的服务范围、客服渠道、退款和取消规则、数据处理边界及合规限制。正式付费前，我们会用邮件确认项目范围、交付物、时间表和报价。

The following information describes data2paper's service scope, support contact, refund and cancellation rules, data handling boundaries, and compliance restrictions. Before any paid engagement, we confirm the scope, deliverables, timeline, and quote by email.

商家名称与服务描述

Business Name & Services

商家名称：data2paper。我们为临床医生、青年 PI、研究生和科研团队提供科研支持服务，包括数据探查与清洗、统计分析、图表生成、论文初稿支持、参考文献核验、可复现代码整理和投稿材料准备。data2paper 不承诺期刊接收或发表结果，所有研究设计、数据来源、临床解释、署名和最终投稿决定均由研究者负责。

Business name: data2paper. We provide research support services for clinicians, junior PIs, graduate researchers, and research teams, including data profiling and cleaning, statistical analysis, figure generation, manuscript draft support, reference verification, reproducible code preparation, and submission package assembly. data2paper does not guarantee journal acceptance or publication. Research design, data provenance, clinical interpretation, authorship, and final submission decisions remain the researcher's responsibility.

客服联系方式

Customer Support

客服邮箱：[email protected]。请在邮件中写明样本量、研究类型、目标期刊或影响因子区间、预期交付物和时间要求。我们通常在周一至周五 9:00-18:00 CST 响应，并尽量在 24 小时内回复。

Support email: [email protected]. Please include sample size, study design, target journal or impact factor range, expected deliverables, and timeline requirements. We usually respond Monday-Friday, 9:00-18:00 CST, and aim to reply within 24 hours.

价格、交付与实物退货

Pricing, Delivery & Returns

免费评估阶段不收费。付费服务仅在双方通过邮件确认书面报价、服务范围、里程碑、交付格式和时间表后开始。data2paper 销售的是数字化科研支持服务，不销售实体商品，因此不适用实物退货或物流退换流程。

The free assessment stage is not charged. Paid service begins only after both parties confirm the written quote, service scope, milestones, delivery formats, and timeline by email. data2paper sells digital research support services, not physical goods, so physical return shipping or return logistics do not apply.

退款与争议政策

Refund & Dispute Policy

项目尚未开始前取消：可全额退款。
项目已经开始但未交付最终成果：按已完成里程碑、已投入人工和已交付材料扣除相应费用，剩余部分退款。
最终交付物已发送并被客户接受后：除非交付物明显不符合已确认的书面范围，一般不再退款。
争议处理：请通过客服邮箱提交订单信息、争议点和相关证据，我们会先尝试通过修改、补充说明或部分退款解决。

Cancellation before work starts: full refund.
Cancellation after work starts but before final delivery: refund is prorated after deducting completed milestones, labor already performed, and materials already delivered.
After final deliverables are sent and accepted: refunds are generally not available unless the deliverables materially fail to match the confirmed written scope.
Disputes: email support with order details, disputed items, and supporting evidence. We will first try to resolve the issue through revisions, clarification, or partial refund where appropriate.

取消政策

Cancellation Policy

客户可以在项目开始前随时通过邮件取消。项目开始后如需取消，请尽快邮件通知；我们会根据已确认里程碑、已完成工作和不可追回人工成本结算。订阅或连续服务（如适用）可在下一计费周期前通过邮件取消。

Customers may cancel by email at any time before work starts. After work starts, please notify us by email as soon as possible; settlement will be based on confirmed milestones, completed work, and non-recoverable labor cost. Recurring services, if any, may be cancelled by email before the next billing cycle.

法律、学术与出口限制

Legal, Academic & Export Restrictions

我们不提供临床诊断、治疗建议、患者护理决策、伦理审批替代服务、数据造假、引用造假、抄袭、署名规避、一稿多投规避或保证发表服务。客户不得上传其无权处理的数据、未脱敏敏感个人信息、受出口管制或机构禁止外传的数据。客户需确保数据使用、伦理审批、知情同意、机构政策和当地法律合规。

We do not provide clinical diagnosis, treatment advice, patient care decisions, substitutes for ethics approval, fabricated data, fabricated citations, plagiarism, authorship circumvention, duplicate-submission circumvention, or guaranteed publication services. Customers must not submit data they are not authorized to process, non-de-identified sensitive personal data, export-controlled data, or data prohibited from external transfer by their institution. Customers are responsible for compliance with data use rights, ethics approval, informed consent, institutional policies, and applicable law.

数据隐私与保密

Data Privacy & Confidentiality

评估阶段仅要求元数据问卷，不要求提交原始数据。正式合作需要原始数据时，请尽量先脱敏。我们使用加密传输、本地分析和最小必要访问原则，不将客户原始数据上传到第三方公开服务。交付完成后，数据按书面约定删除或返还。

During assessment, we only request metadata and do not require raw data. When raw data is required for a paid engagement, please de-identify it where possible. We use encrypted transfer, local analysis, and least-necessary access, and we do not upload customer raw data to third-party public services. After delivery, data is deleted or returned according to the written agreement.

促销条款

Promotion Terms

免费评估仅用于判断项目可行性、风险和工作量，不等同于正式交付物、现金折扣或可转让权益。任何限时折扣、积分或优惠券（如适用）必须以邮件或页面上的书面条款为准，不能兑换现金，不可转让，且不能与未明确说明可叠加的其他优惠同时使用。

The free assessment is used only to evaluate feasibility, risk, and workload. It is not a formal deliverable, cash discount, or transferable benefit. Any time-limited discount, credit, or coupon, if applicable, is governed by the written terms shown by email or on the page. Promotions have no cash value, are non-transferable, and cannot be combined with other offers unless explicitly stated.

从临床数据到可审核论文初稿

From Clinical Data to a Review-Ready Draft