DATA SWEEP helps you search scientific literature, generate AI-powered literature reviews, inspect cited papers, export results, and organize research in projects.

How it works?

1. Enter your research question or topic.
2. Optionally add filters to narrow the articles included in the search.
3. Start the search.
4. DATA SWEEP creates a Chat with an AI-generated literature review and the list of cited articles.
5. From there, inspect articles, run bibliometric analysis, export, or save everything to a project.

---

This guide covers:

• How to search
• Search modes: Basic, Pro, and Deep
• Chat
• Bibliometric analysis
• Exporting articles and the literature review
• Article page
• Projects
• Agreement Meter
• Best practices
• Quick FAQ

---

How to Search

Type a research question, topic, or scientific claim in the search box, then start the search. Example searches:

• How does virtual reality tutoring affect STEM education?
• What is the impact of venture capital on startup growth?
• Can biochar amendments improve drought tolerance in maize?

Query quality

DATA SWEEP checks whether your query is suitable for a scientific literature search.

A query can be treated as:

• Perfect: clear, complete, and structured enough to search directly.
• Good: valid, but broad or missing detail. DATA SWEEP may suggest improved versions.
• Bad: not recognizable as scientific research, or not suitable for search.

Search Modes

Not sure which mode to pick?

• Need a quick answer → Basic
• Starting a new topic or writing a standard review → Pro (recommended for most searches)
• Writing a thesis, systematic review, or comprehensive report → Deep

Mode	Papers analyzed	Output	Best for
Basic	Up to 10	Concise overview	Quick answers and short summaries
Pro	Up to 30	Overview + 3-column comparison table	Most searches (default)
Deep	Up to 100	Overview + in depth comparison table	Comprehensive literature reviews

Deep runs in two passes: first building an outline with themes, clusters, and gaps, then writing the full review. The table covers the background, purpose, methods, results, conclusions, limitations, and future directions of each analyzed paper.

Language detection: DATA SWEEP automatically detects the language of your query and defaults to English if uncertain. English queries often return the broadest coverage since most academic databases index in English.

Filters

Filters help you narrow the search results.

Filter	What it does	Example
Publication year	Limit results to a year range	2018–2024
Minimum citations	Show papers with at least N citations	10, 50, 100
Open access	Show only open access papers
Journal rank	Filter by quartile	Q1, Q2
Journal name	Search within selected journals	Nature, PLOS ONE
Academic domain	Filter by research field	Computer Science, Medicine
Paper language	Choose the language of included papers	English, Romanian

Applying filters

Apply your filters before starting the search.

Use filters carefully. Too many filters can remove useful papers, especially for narrow topics.

Changing filters after a search

You can change filters directly from the chat. When you regenerate after changing filters, DATA SWEEP searches again and creates a new literature review using the updated set of articles.

---

Chat

A chat is the workspace where your research lives. When you run a search, DATA SWEEP creates a chat and adds the literature review and cited articles as the first message. Each time you search again within the same chat, a new message is added, so you can run multiple searches on a topic and keep everything in one place.

A chat can be named, saved to a project, and shared.

Each message in a chat includes:

• Your original query.
• The generated literature review.
• The answer language and text style.
• The cited articles.

What you can do from a chat

• Read the AI-generated literature review.
• Browse cited articles and load up to 500 more.
• Adjust the language, text style, and citation format.
• Change filters and regenerate the answer with a different set of papers.
• Cite individual articles.
• Save articles or the whole chat to a project.
• Run a bibliometric analysis.
• Export the literature review or the article list.

How summaries are generated: The literature review is built from your research question and the titles and abstracts of the retrieved articles. A more specific query and more relevant articles generally produce a more focused and accurate review.

Chat Controls

These controls let you refine the answer within a chat without running a new search.

Change text style

Switch the summary between Bullet Points and Paragraphs. The meaning and citation markers are preserved.

Change language

Translate the answer into a supported language. Citation markers, text formatting, and academic tone are preserved throughout.

Change filters

Changing filters triggers a new search and regenerates the literature review. It counts against your search quota.

Results Section

Below the literature review, DATA SWEEP shows the cited articles and additional papers beyond the initial cited set.

Each article card can include:

• Article title.
• Generated TLDR.
• Publication year.
• Citation count.
• Authors.
• Journal or conference name.
• Open access status.
• PDF, source, or attribution links when available.
• Citation button.
• Add to project button.

Cite an article

Click the citation icon to open the citation window.

Available citation formats include:

• BibTeX
• APA
• MLA
• Chicago
• Harvard
• IEEE

After selecting a format, copy the citation for use in papers, reports, notes, or reference managers.

Add an article to a project

Click the plus button to add an article to a project.

Projects help you save and organize useful papers for later.

---

Bibliometric Analysis

Click Bibliometrics in the chat to generate a bibliometric report for your search.

The analysis runs a similarity-based search across up to 10,000 articles and compiles a structured PDF report.

The report is divided into three areas:

Overview

1. General Statistics: total articles analyzed, year range, total citations, average citations, median citations, h-index, and reference counts.
2. Temporal Analysis: publication trend over time and total citations per year.
3. Top Cited Articles: the 10 most cited papers with authors, years, and DOI links.

Journal and author analysis

1. Journal Analysis: top 10 journals by number of publications.
2. Quartile Distribution: share of articles published across Q1, Q2, Q3, Q4, and unindexed journals.
3. Author Productivity: top 10 most productive authors and their individual papers.
4. Collaboration Analysis: average authors per paper, degree of collaboration, and single-author paper count.

Content and access

1. Open Access Analysis: percentage and breakdown of open access articles.
2. Citation Distribution: histogram showing how citations are distributed across the result set.
3. Title Keyword Analysis: top 20 most frequent words in article titles.

Bibliometrics provides a broader landscape view of a topic than the literature review alone. It should be used alongside, not instead of, reading the articles.

---

Exporting Articles and Chats

DATA SWEEP lets you export the article list and the full chat.

You can export:

• The full chat as PDF.
• A selected set of articles by current view or by range, up to 5,000 articles depending on your plan.

Method	Use when
Currently loaded	You want the articles already visible on screen
Range	You want a specific slice of the ranked list, such as the top 100

Export currently loaded articles

Use Currently loaded to download the articles already visible on the page.

1. Click the Results near the article list.
2. Select Currently loaded.
3. Choose the format.
4. Choose whether to receive the file by email.
5. Start.

Export a range of articles

Use Range to download articles from a specific position in the ranked list.

1. Click Results.
2. Select Range.
3. Enter the starting position.
4. Enter the ending position.
5. Choose the format.
6. Choose whether to receive the file by email.
7. Start.

Export formats

Format	Use case	Plan
CSV	Spreadsheet analysis	All plans
Excel	Spreadsheet analysis	Premium and above
BibTeX	Reference managers (LaTeX, Zotero)	Premium and above
RIS	Reference managers (Mendeley, EndNote)	Premium and above

Export limits

Your subscription plan controls:

• Maximum articles per export.
• Available export formats.

If the selected export is larger than your plan allows, reduce the range or export fewer loaded articles.

---

Article Page

Open an article to inspect a source in detail before citing or saving it.

Beyond the article card, the Article page shows the full abstract, DOI link, source link, and tabs for the article's citations and references.

Citations and references

Two tabs appear at the middle of the page:

• Citations: papers that cite this article. Useful for finding newer research that builds on it.
• References: papers this article cites. Useful for tracing the background and source literature.

You can download both citations and references.

---

Projects

Projects help you organize your research across multiple sessions. Access your projects from the main navigation.

You can add items to a project in two ways:

• Save an entire chat using the save button in the chat.
• Add individual articles using the plus button on any article card.

From a project, you can browse saved chats and articles and export them at any time.

---

Agreement Meter

Agreement Meter is for testing a scientific claim against article evidence.

Use it when your input is a claim, not a question.

Good examples:

• Daily omega-3 supplementation lowers systolic blood pressure in adults with hypertension.
• Chronic exposure to airborne microplastics increases systemic inflammation markers in humans.

Agreement Meter scores each cited article as agreement, disagreement, or neutrality toward the claim, with a short rationale and confidence-style percentages.

Example output:

Article	Stance	Confidence	Rationale
D. Alexander et al. (2014)	Agreement	90%	This meta-analysis including 70 randomized controlled trials found that supplementation with EPA and DHA significantly reduced systolic blood pressure by an average of 1.52 mm Hg compared to placebo...
Liyu Tao et al. (2020)	Neutrality	80%	The abstract describes a protocol for a systematic review that aims to evaluate the effects of omega-3 supplementation on blood pressure in hypertensive patients, focusing on systolic and diastolic blood pressure as primary outcomes. However, it does not provide any actual results or conclusions about whether omega-3 effectively lowers blood pressure...
C. Byrne et al. (2020)	Disagreement	85%	The study found no significant difference in median triglyceride concentrations between omega-3 users and non-users, and importantly, omega-3 fatty acid use was not associated with a reduction in cardiovascular events or death...

---

Best Practices

Use the right mode

What you need	Mode
Quick overview	Basic (10 papers)
Structured synthesis with comparison table	Pro: default (30 papers)
Comprehensive review with full study breakdown	Deep (100 papers)
Testing a claim against the evidence	Agreement Meter

Start broad, then narrow

Run the first search without filters. Use the initial output to understand the research landscape, then add filters step by step.

Too many filters at once can remove useful papers, especially on narrow topics where the literature is limited. If results become too narrow or irrelevant, remove some filters and regenerate.

Check the articles

The AI summary helps you understand the topic faster, but you should still check the articles.

Open the most relevant articles and read their abstracts before using them in your work. For important work, read the full paper.

Treat summaries as evidence maps

DATA SWEEP summaries are best used as maps of the available evidence, not as final conclusions.

Check:

• Whether the cited papers actually support each claim.
• Whether important populations or methods are missing.
• Whether the most relevant articles are open access or require a subscription.

Use citations

When using information from a paper, cite the source.

DATA SWEEP helps you copy citations in several formats.

Use projects for longer work

If you are researching over several days, save useful chats and articles into projects.

This helps you avoid losing important sources.

---

Quick FAQ

What should I do if results are not relevant?

Try one or more of the following:

• Rewrite the question more clearly.
• Use one of the suggested query improvements.
• Add context such as population, method, outcome, location, or time period.
• Use filters.
• Change the search mode.
• Search in English for broader academic coverage.

Why did Agreement Meter reject my question?

Agreement Meter expects a claim, not a question.

Use:

Omega-3 supplementation reduces cardiovascular risk in adults over 50.

Not:

Does omega-3 supplementation reduce cardiovascular risk in adults over 50?

Should I trust the AI summary completely?

No. The summary is useful for understanding a topic faster, but important information should always be checked against the original articles.

Why are some articles not used in the summary?

DATA SWEEP instructs the AI to exclude articles when their title or abstract does not provide enough information to support the answer.

Why are citation numbers used inside the summary?

Citation numbers connect each claim in the summary to the article list. They make it easier to verify which papers support each statement.

Why did changing filters create a new answer?

Filters change the article set. When the evidence set changes, DATA SWEEP must regenerate the answer so the literature review reflects the updated article set.