How to Use Logseq for Research ft. Cara Antonaccio

Logseq’s biggest payoff for research isn’t just note-taking—it’s how a daily journal becomes a living research index through automatic linking. Cara Antonaccio runs her mornings through a Journal page where every scheduled item and research activity is logged as a “process,” and those lines inherit links to the topics she’s tracking (e.g., armed conflict, mental health problems, conflict-affected children). The result is a continuously updated web of references: when she writes under a dated entry, the content is already connected to the relevant tags and concepts, making later navigation, querying, and cross-reading far easier than when notes sit in isolated folders.

Her research workflow is built around the same principle: structure first, then let relationships emerge. Antonaccio studies armed conflict and uses algorithmic modeling and geographic statistics, and she keeps her dissertation work in Logseq rather than treating it as a separate “writing tool.” She starts with a tightly structured outline for each source—copying the article’s headers and writing her own notes under them—so she synthesizes ideas in her own words instead of highlighting and pasting. That approach also functions as a plagiarism safeguard because the notes are processed, not copied.

Sources enter her system through a deliberate pipeline. She builds a literature library where each article or book gets its own row and a direct link to the source, then uses Logseq’s built-in outline extraction to create a consistent starting structure. For discovery, she relies on Google Scholar and Connected Papers. She avoids Zotero during active writing because it can blur what she has already decided to cite versus what she’s still exploring; instead, she keeps citations accessible in the draft and only switches to better BibTeX naming conventions later, once arguments stabilize.

When she needs to understand the literature’s shape, Connected Papers helps map “seminal works” and their derivative or citing papers, even exporting results to CSV for quick analysis. As her dissertation advances, she drafts with the same outlining discipline—headings first, then one idea per block—while embedding math via LaTeX and inserting figures and diagrams imported from Lucidchart (exported as SVG). For coding and analysis planning, she keeps R scripts and even GitHub README content inside Logseq so she can reference algorithms while writing.

Antonaccio’s workflow also stays intentionally non-linear: searching, outlining, reading, modeling, drafting, and editing run in parallel across multiple projects. She uses Logseq’s graph view to see how authors, tags, and citations cluster—her dissertation’s random forest method becomes a visible hub surrounded by related articles. Exporting is workable but not perfect; she prefers copying HTML into Word or Google Docs so heading levels map cleanly, and she adjusts formatting when needed (including dealing with bullet artifacts). Overall, her core lesson is practical: start simple, keep the system consistent, and integrate planning, research, and writing so the structure of the work can grow from the daily record rather than being imposed later.