.. _literature_survey: Literature survey =================== A literature survey, or literature review, is where you investigate whether other people have tried to solve the same problem as you and, if so, how they did it. The outcome of the literature survey is not only a synthesis of different solution methods in the literature but also identification of gaps in the state of the art. Your research should fall within said gaps. The information here is terse. I highly recommend you google to find more guides. Also, go talk to the librarians! It's literally their job to do this. Overview of steps ------------------- `This article `_ has a nice summary of the process. Some extra tips from me are below. Start with Scite AI ------------------------------- * DTU has access to Scite AI, which is a good tool to get started. * **!!! BUT ASSUME THE TOOL HALLUCINATES !!!** Use Scite AI to figure out which papers might be relevant, but then read the papers yourself and draw your own conclusions. * Get started with Scite AI: * Go to `DTU Findit `_ and log in. * Click the "Scite AI" link on "Other search tools", right above the search box. * Create an account in Scite AI using your DTU credentials and verify your email. * Use the "Assistant" chatbot to search literature for various topics. Explore the other tabs at your leisure. * If you can't get the paper through Scite AI, use DTU Findit to get the paper PDF. Keeping track of papers ----------------------------- * `Zotero `_ seems to be the best current solution for tracking papers as you read them. It also integrates nicely with Overleaf. * You can use tags/collections in Zotero to indicate if you've read the paper yet, which research question it's relevant for, etc. * I like making a table with all the papers I read. Each row is a new paper. The columns are: BibTeX ID, title, 1-sentence summary, categories (e.g., which research questions it's relevant for, what topic does it fall under, etc.), key conclusions from their results (bullet list) pros (bullet list), cons (bullet list) and notes (bullet list). Spar with your supervisor ----------------------------- * I like my students to present their analysis of, e.g., 2-3 papers in supervision meetings. It's a nice way for me to not only keep updated on relevant literature but also verify the student's analysis/interpretation of the papers. * See example slide in PPTX template under "Weekly meetings" in :ref:`project_management` Synthesizing the results ------------------------- * I recommend using a synthesis matrix to synthesize the information across different papers by topic. Here are some examples: * https://guides.library.jhu.edu/ld.php?content_id=16149009 (my preference) * https://www.concordia.ca/content/dam/library/docs/research-guides/gradproskills/Lit-review-synthesis-matrix-Word.docx I recommend the first example because each row corresponds to a topic, which should naturally become a paragraph (or more) when you write the actual literature survey. Writing the literature survey ------------------------------ Documentation of the literature survey becomes a "State of the art" section in your thesis. This section should answer two questions: 1. How have other researchers tried to solve your problem? 2. Where are the gaps in the state of the art? In engineering, literature surveys are organized by topic/methodology. It is common to subdivide state-of-the-art sections into subsections by topic for easier reading. Papers may be cited in multiple sections if they span topics. Examples of literature surveys ------------------------------- * In examples of project plans: :ref:`project_plan`. * Example review paper: https://iopscience.iop.org/article/10.1088/1742-6596/2362/1/012035/meta.