# How to Organize Dissertation Figures and Tables By the end of this guide you will have a working system for organizing dissertation figures and tables from the first draft to the final submission. You will build a master log that tracks every visual item, set up a numbering scheme that survives chapter reorganization, generate captions and cross-references automatically, and run a structured quality check before you submit. The whole system takes about two hours to set up and saves weeks of renumbering pain later. You will need one of the following tool sets on hand: - Microsoft Word or Google Docs for the word-processing route - A LaTeX distribution such as TeX Live or Overdose (Overleaf) for the typesetting route - Python 3 with the pandas and matplotlib libraries installed, or Excel or [Google Sheets if](/blog/data-analysis/google-sheets-if-function) you prefer a spreadsheet for the log - Your dissertation outline, your data files, and a folder where all figure exports will live None of these tools cost money. Python, pandas, and matplotlib are free. Overleaf has a free tier. Word is usually provided through your university. ## Why Figure and Table Organization Matters for Dissertation Planning A dissertation can contain forty or more figures and tables spread across five or more chapters. When those items are unorganized, three predictable failures occur. Numbers get reused after chapters are reordered. Captions drift out of sync with the data. Reviewers find figures that the text never mentions. Organizing dissertation figures and tables early is also a practical strategy for how to finish a dissertation. A maintained log tells you exactly which items are done, which are placeholders, and which data analyses are still missing. That visibility converts a vague task, "do the figures," into a finite checklist you can complete chapter by chapter. Well-organized visuals also carry real communicative weight. In one informetric study of Chinese nursing dissertations, researchers retrieved all doctoral dissertations and master's theses tagged "nursing" in the CNKI database, ran a co-word analysis on high-frequency keywords, and plotted a single co-occurrence figure that captured nine distinct research topics at a glance [1]. One organized figure summarized the structure of an entire research field. That is the standard of compression your dissertation visuals should aim for. ## Figure or Table: Choosing the Right Format ![Decision path asking whether exact values or the pattern matter, leading to a table or a figure, with a note against showing both.](/article-images/illustrations/organizing-dissertation-figures-and-tables-figure-or-table-choice.webp) Before you organize anything, decide what belongs in a figure and what belongs in a table. The decision rule is simple. Use a figure when the pattern matters more than the exact values. Use a table when the exact values matter more than the pattern. | Situation | Best format | Reason | |--|--|--| | Showing a trend over time or dose | Line figure | Readers see the shape of the relationship instantly | | Comparing a few group means with error bars | Bar or dot figure | Differences in magnitude are visually obvious | | Reporting exact coefficients, [confidence intervals](/blog/data-analysis/confidence-interval-formula-examples), p values | Table | Readers need precise numbers to check your claims | | Presenting a conceptual model or framework | Figure (boxes and arrows) | Structure and direction of influence are the message | | Showing participant flow through a study | Figure (flow diagram) | Attrition at each stage is easier to trace visually | | Listing demographic characteristics of a sample | Table | Many variables, each with its own value, no pattern to show | | Displaying a spatial or network relationship | Figure | Position carries meaning that rows cannot | A useful corollary: never present the same data as both a detailed figure and a full table. Pick one, and mention the other option in text only if a reader would genuinely need it. Duplication inflates page counts and invites inconsistency between the two versions. ## Organizing Dissertation Figures and Tables Step by Step This walkthrough gives you a complete pipeline. Steps 1 through 3 apply regardless of software. Steps 4 through 6 differ between LaTeX and Word, and both routes are shown. ### Step 1: Build a Master Log of Every Visual Item ![Flow from making the spreadsheet through item IDs, source data, status values and a log update at the end of each session.](/article-images/illustrations/organizing-dissertation-figures-and-tables-master-log-setup.webp) Create a spreadsheet, or run this Python snippet to generate the same thing as a CSV file: ```python import pandas as pd columns = [ "item_id", "type", "chapter", "caption_draft", "source_data", "file_name", "status", ] log = pd.DataFrame(columns=columns) log.to_csv("figure_table_log.csv", index=False) print(log.shape) ``` Expected output: ``` (0, 7) ``` The log starts empty with seven columns. Fill it with one row per figure or table. Here is what a populated log looks like after an hour of inventory work: ```python rows = [ ["F2.1", "figure", "2", "Study design overview", "design_notes.docx", "fig2_1_design.png", "done"], ["F3.1", "figure", "3", "Mean response by treatment group", "exp1_raw.csv", "fig3_1_means.png", "done"], ["T4.1", "table", "4", "Regression model coefficients", "model1_summary.csv", "tab4_1_coeffs.csv", "draft"], ["F5.1", "figure", "5", "Participant flow through the study", "screening_log.xlsx", "fig5_1_flow.png", "todo"], ] log = pd.DataFrame(rows, columns=columns) log.to_csv("figure_table_log.csv", index=False) print(log.to_string(index=False)) ``` Expected output: ``` item_id type chapter caption_draft source_data file_name status F2.1 figure 2 Study design overview design_notes.docx fig2_1_design.png done F3.1 figure 3 Mean response by treatment group exp1_raw.csv fig3_1_means.png done T4.1 table 4 Regression model coefficients model1_summary.csv tab4_1_coeffs.csv draft F5.1 figure 5 Participant flow through the study screening_log.xlsx fig5_1_flow.png todo ``` The columns that matter and why: - **item_id** combines a type prefix (F for figure, T for table) with the chapter number and a sequence number. This ID appears in your file names, your log, and your drafts, so every reference to an item matches everywhere. - **source_data** records exactly which file the item was built from. When a reviewer questions a number two years from now, you find the origin in seconds. - **status** takes one of three values: done, draft, or todo. At any moment the log tells you how much visual work remains, which is the core of realistic dissertation planning. ### Step 2: Map Every Item to a Chapter and Check the Balance Sort the log by chapter and count items per chapter: ```python counts = log.groupby("chapter")["item_id"].count() print(counts) ``` Expected output: ``` chapter 2 1 3 1 4 1 5 1 Name: item_id, dtype: int64 ``` Two balance checks are worth doing at this stage. First, a results chapter with zero figures and ten tables of raw output usually means you have not done the analytical work of deciding what the reader needs to see. Second, a discussion chapter with no visuals at all is often correct, but if your argument depends on comparing two patterns, one summary figure there can carry the section. The goal is not a fixed number per chapter. The goal is that every chapter's visuals match what that chapter actually argues. ### Step 3: Set a Numbering and Caption System Use chapter-based numbering. In chapter 3, figures are Figure 3.1, Figure 3.2, and so on. Tables restart independently as Table 3.1, Table 3.2. The formula is: ``` item number = chapter number . position of item within that chapter's type ``` This scheme survives reorganization. If you delete Figure 3.4, only later items in chapter 3 shift, and your log makes the shift mechanical. Sequential numbering across the whole document (Figure 1 through Figure 30) breaks the moment a committee asks you to move a section. Write captions as descriptive sentences, not fragment titles. Compare these two: - Weak: "Group means" - Strong: "Mean response concentration by treatment group after 14 days, with 95 percent confidence intervals" The strong caption answers what, under what conditions, and what the variability measure is, without requiring the reader to hunt through the text. A caption should let a reader who only looks at figures and tables still follow the main argument of the dissertation. Many committee members read exactly that way on a first pass. ### Step 4: Produce Figures From a Single Style Template Every figure in the dissertation should share font, line weight, color palette, and resolution. The cheapest way to guarantee this is one template script that produces all data figures. Here is a matplotlib template: ```python import matplotlib.pyplot as plt plt.rcParams.update({ "font.size": 10, "font.family": "sans-serif", "axes.spines.top": False, "axes.spines.right": False, "figure.dpi": 300, "savefig.dpi": 300, }) groups = ["Control", "Low dose", "High dose"] means = [12.4, 18.1, 25.7] fig, ax = plt.subplots(figsize=(6, 4)) ax.bar(groups, means, color="#4C72B0") ax.set_xlabel("Treatment group") ax.set_ylabel("Mean response (mg/L)") fig.savefig("fig3_1_means.png", bbox_inches="tight") ``` Expected output: a PNG file named `fig3_1_means.png`, 6 by 4 inches at 300 dots per inch, with three bars, no top or right border lines, and axis labels in a 10 point sans-serif font. The values here are placeholder example data for the template, not findings from any study. Key parameters and what they control: - **dpi 300**: most university submission guidelines and print shops expect at least 300 dots per inch for raster images. Lower values look blurry when printed. - **figsize (6, 4)**: sets the physical size the figure occupies at 100 percent scale in your document. Match this to your page's text width so you never stretch a small image. - **bbox_inches="tight"**: trims surrounding white space so the figure does not arrive with an invisible margin that throws off your layout. - **spines off (top and right)**: removes chart junk. Most scientific style guides favor clean axes over boxed plots. Export figures as PNG for raster plots, PDF or SVG for vector graphics if your submission system accepts them, and always keep the script that generated the figure alongside the export. A figure you cannot regenerate is a figure you cannot fix. For tables, the equivalent discipline is consistent formatting: same number of decimal places within a column, units in the column header rather than repeated in every cell, and significance markers defined in a footnote under the table. ### Step 5: Insert Items and Cross-Reference Them Automatically Manual numbering is the single most common source of dissertation errors. Automate it. In LaTeX, insert a figure like this: ```latex \begin{figure}[htbp] \centering \includegraphics[width=0.8\textwidth]{figures/fig3_1_means.png} \caption{Mean response concentration by treatment group after 14 days.} \label{fig:group-means} \end{figure} ``` Then reference it in the text by label, never by typed number: ```latex Response increased across dose groups, as Figure~\ref{fig:group-means} reports. ``` The `htbp` placement specifier asks LaTeX to place the figure here, at the top of a page, at the bottom, or on a float page, in that order of preference. The `\label` command must come after `\caption` inside the figure environment or the reference will point at the wrong number. When you reorganize chapters and recompile, every `\ref` updates automatically. In Word, the equivalent workflow is: 1. Select the image, open the References tab, choose Insert Caption. Pick the Label "Figure" and check "Include chapter number" if your headings use multilevel numbering. 2. Type the descriptive caption text after the auto-generated number. 3. In the body text, choose References, then Cross-reference, select "Figure" and "Only label and number," and click the item you mean. 4. Press Ctrl+A then F9 before submission to update every field at once. The most common Word failure mode is a document that prints "Figure 0.1" or "Error! Reference source not found." That error message appears when a cross-reference points to a caption that was deleted or pasted without its bookmark field. Fix it by reinserting the caption on the target item and pressing F9 to refresh fields. ### Step 6: Generate the List of Figures and List of Tables Most dissertations require a List of Figures and a List of Tables after the table of contents. Because you used caption fields or LaTeX captions, these lists generate themselves. In LaTeX, place this where the lists belong: ```latex \listoffigures \listoftables ``` After compiling twice, the output contains every caption with its page number, formatted consistently. If an item is missing from the list, the cause is almost always a caption typed as plain text rather than through the `\caption` command, or in Word, a caption not applied with the Caption style. In Word, place the cursor where the list goes, choose References, then Insert Table of Figures, and select "Figure" as the caption label. Repeat with "Table." ### Step 7: Run a Final Quality Check Before submission, verify each of these against the log: 1. Every item in the log appears in the document, and every item in the document appears in the log. 2. Every figure and table is cited in the text at least once, in numerical order on first mention. Figure 3.2 should not be discussed before Figure 3.1. 3. Numbers in the text match the numbers in the item. Recompute two or three key values from `source_data` as a spot check. 4. Captions are self-contained sentences with units and the variability measure. 5. All axes have labels with units, all table columns have headers with units. 6. The List of Figures and List of Tables match the actual items and page numbers. 7. Resolution is at least 300 dpi for every raster image, verified by right-clicking the file properties or checking the export settings. Print the checklist or turn it into a column in your log. Checking forty items from memory is how mistakes survive to the defense. ## Worked Examples From the Research Literature Three examples from published work show these principles in action, one for each major item type. **A flow diagram.** A 2024 scoping review protocol on clinical incident analysis methods in acute care settings states that the review follows JBI methodology and PRISMA-ScR reporting guidance [2]. PRISMA-ScR, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews, requires a flow diagram documenting how records were identified, screened, and included. If your dissertation includes a literature review of this kind, the same flow diagram is the standard visual for your methods chapter, and following its published reporting structure signals methodological rigor to examiners. **A conceptual framework figure.** Researchers developing health communication materials for adolescent technology addiction built their intervention around a "4C" risk model (Control, Craving, Compulsion, Consequences) and a "3A" response strategy (Ask, Acknowledge, Allow), refining the materials through a six-step iterative process with expert validation and pretrial testing [3]. A conceptual figure with boxes for each 4C domain and arrows mapping each domain to the corresponding 3A response is exactly the kind of framework diagram that earns a dedicated early figure in a dissertation. Notice also how the team reported their validation outcome systematically, with over 90 percent of pretrial respondents giving positive ratings on clarity, relevance, visual appeal, and cultural appropriateness [3]. When your dissertation includes an evaluation like that, a results table with the criterion rows and rating proportions is the cleanest presentation. **A synthesis figure.** The co-word analysis of nursing dissertations mentioned earlier plotted a co-occurrence figure of high-frequency keywords that represented nine research topics, giving health staff a map of hot topics in the field [1]. If your dissertation involves bibliometric or thematic analysis, this is your model: one figure that turns a large corpus into an interpretable structure, placed at the point where the reader needs the overview rather than buried in an appendix. ## Organizing Dissertation Figures and Tables Tips These are the practices that separate a functional system from a fragile one: 1. **Name files by item ID, never by content whim.** `fig3_1_means.png` sorts correctly next to `fig3_2_flow.png`. `final_final_REAL_means.png` sorts nowhere usefully. 2. **Version the script, not the export.** Keep `plot_figures.py` under version control if you can. If a reviewer asks for a logarithmic y axis, you change two lines and re-export every figure consistently. 3. **Keep one directory per chapter**, mirroring your document structure: `figures/ch3/`, `figures/ch4/`. A flat folder of thirty files becomes unnavigable fast. 4. **Use a colorblind-safe palette.** The default matplotlib palette and palettes such as ColorBrewer's qualitative sets are designed to remain distinguishable under common color vision deficiencies. Add direct labels or distinct marker shapes so the figure survives grayscale printing. 5. **Round consistently.** If means are reported to one decimal place in a table, the same variable is reported to one decimal place everywhere, including the text. 6. **Schedule the log update.** Add "update the figure log" to the end of each writing session. A log that is three chapters stale is worse than no log, because you will trust it. 7. **Do the placeholder pass early.** Even before analyses are complete, insert every planned item with a draft caption and a status of todo. This front-loads the structure and gives your advisor something concrete to react to, which accelerates how to finish a dissertation rather than delaying it. 8. **Check your graduate school's specific formatting manual.** Programs differ on caption placement above or below, on whether tables count toward page limits, and on required margins for landscape tables. The handbook overrides every general rule in this article. ## Common Mistakes and Limitations These mistakes account for most of the corrections examiners and format editors request: - **Figures never cited in the text.** An orphan figure invites the question of why it exists. Every item earns its page space by being discussed. - **Duplicate presentation.** The same data as a full table and a full figure, with the two slightly disagreeing because one was updated after the other. Present each dataset once. - **Screenshots of low resolution.** A screenshot of a statistical software output pasted at screen resolution looks acceptable on a laptop and unacceptable in a bound copy. Rebuild the table natively or export at print resolution. - **Caption fragments.** "Results of the survey" tells the reader nothing. A caption is a sentence that survives being read alone. - **Missing units.** A y axis labeled "Concentration" without units, or a table column headed "Weight" without kilograms or pounds, forces every reader to reconstruct context from the methods chapter. - **Inconsistent decimal places** within a single column, which suggests careless assembly and undermines confidence in the underlying numbers. - **Last-minute assembly.** Building all figures in the final month guarantees inconsistency because early figures and late figures are made under different conditions and, often, with different script versions. This is the avoidable version of the end-of-dissertation crunch. - **Broken cross-references submitted as-is.** Word documents with "Error! Reference source not found" have gone through real defenses. Press F9 and re-read the field results before every submission. Two limitations apply to the system itself. First, your institution's thesis and dissertation manual is the final authority on formatting details, and where it conflicts with general practice described here, follow the manual. Second, individual projects differ in what examiners expect for visual density, so treat the per-chapter balance checks as prompts for a conversation with your advisor rather than as targets with fixed thresholds. ## Frequently Asked Questions ### How many figures and tables should a dissertation have? There is no universal number. Include every item that advances the argument and earns a citation in the text, and cut anything that exists only for completeness. A typical empirical dissertation carries somewhere between fifteen and forty items, but your discipline's norms and your committee's preferences matter more than any range. ### Do figures and tables count toward the page limit? This depends entirely on your graduate school's handbook. Some programs count everything, others exclude appendices and some front matter. Check the manual before you write, because the answer changes how you allocate space. ### Where should figures and tables be placed in the document? Place each item after its first mention in the text, as close to that mention as the layout allows. Some departments require all figures at the end of the document or in appendices. Follow your program's manual when the two conflict. ### How do I number figures and tables across multiple chapters? Use chapter-based numbering: Figure 3.1 is the first figure in chapter 3, and tables restart independently as Table 3.1. This survives moving sections between chapters far better than continuous numbering through the whole document. ### Can I reuse a figure from a published paper in my dissertation? Often yes, but it depends on the publisher's terms and your university's rules. Many publishers allow reuse of your own work in a dissertation with proper citation, and some require a permission statement or a footnote. Check the copyright terms of the specific paper and your graduate school's policy before including it. ### What should a figure caption include? A descriptive sentence stating what is shown, the conditions or sample it applies to, and the variability measure if one is plotted. Add units and define any abbreviations the figure uses. The caption should be understandable without reading the body text. ### How do I make an automatic list of figures in Word? Apply every figure caption with the References tab's Insert Caption tool, then choose References, Insert Table of Figures, and select Figure as the label. Press Ctrl+A and F9 before submission to update all page numbers and entries. ### What file format should dissertation figures be? Use PNG or TIFF at 300 dpi or higher for raster images, and PDF or SVG for vector graphics if your submission system accepts them. Avoid JPEG for charts and diagrams because its compression blurs text and lines. ## Sources 1. [The Study of Hotspots on Nursing Dissertation Based on Co-Word Analysis in China.](https://pubmed.ncbi.nlm.nih.gov/29857462/) 2. [Exploring approaches to contemporary clinical incident analysis methods within acute care settings: a scoping review protocol.](https://pubmed.ncbi.nlm.nih.gov/38126358/) 3. [Pathway to evolve IEC materials for addressing technology addiction among adolescents: An iterative, participatory, mixed-method approach.](https://pubmed.ncbi.nlm.nih.gov/42923853/)