Document a scatter plot workflow by recording the title, exact x,y rows, separate valid, invalid, and over-limit counts, displayed axis ranges, checked point values, and the downloaded SVG; only the first 200 non-empty input rows are parsed for plotting.
This creates a reproducible trail: another reader can see what was entered, what the tool accepted or skipped, how the axes were scaled, and which exported file was reviewed. In the Scatter Plot Maker, the preview and download come from the same generated SVG string, connecting the on-screen result to the standalone 800 by 500 file.

Define What the Plot Record Must Prove
A useful record covers both the actions you took and the evidence you checked. Keep the source data separate from the visual output: the chart shows relative position, while the original rows remain the authoritative record for exact values and later analysis.
- Purpose: Explain what x and y represent, include their units or meaning when relevant, and state the comparison you intended to make.
- Preparation: Identify the source data and record any sorting, filtering, pairing, or manual correction performed before pasting.
- Generation: Preserve the title and rows submitted to the tool, together with the valid, invalid, and over-limit counts.
- Review: Document the displayed axis ranges, representative point values, visual issues, and the name of the exported SVG.
| Record | What to save | What it establishes |
|---|---|---|
| Source | Exact x,y rows and data origin | What was submitted for plotting |
| Parser summary | Valid, invalid, and over-limit counts | Which rows were plotted, skipped, or left unparsed |
| Scale check | Displayed minimum and maximum for each axis | How numeric values became positions |
| Visual check | Selected point values and any overlap | Whether representative points match the source |
| Output | SVG filename and export date | Which finished artifact was delivered |
| Limitations | Relevant notes about scale and interpretation | What the chart does and does not establish |
Prepare the Rows and Title
Prepare a clean text version before pasting. Every non-empty line must contain exactly two comma-separated finite numbers in x,y order. Signed values, decimals, zero, and scientific notation are accepted. Do not add headers, units, labels, extra commas, or empty components because those rows fail the strict parser. Blank lines themselves are ignored, and negative zero is normalized to zero.
Keep the title as ordinary text and paste the prepared rows into the relevant input. If your numbers begin in two spreadsheet columns, the guide to make a 2-variable scatter plot from Excel columns can help with arranging them into the required row format. Do not silently repair rejected rows; retain the original list so your documentation can identify what was corrected.
| Input condition | Tool behavior | Documentation note |
|---|---|---|
| Valid row within the first 200 non-empty rows | Becomes one blue circle | Keep the corresponding source row |
| Malformed row within the first 200, including a magnitude above 1e12 | Is skipped and counted as invalid | Record the reason and any correction |
| Non-empty row after the first 200 | Is counted as over-limit and is not parsed | Keep it in the source data, but not in the plotted set |
| Blank line | Is ignored | It does not count as valid, invalid, or over-limit |
| Entire input above 50,000 UTF-16 code units | Is rejected | Reduce the input before attempting generation |
| No valid point remains | Generation reports an error and clears any older chart | Correct the source and generate again |
Create the Plot and Capture Each Action
Use this sequence to create a plot and document its exact generation state.
- Preserve the source. Save a copy of the title and prepared rows before pasting them. This becomes the authoritative input record if the preview needs to be regenerated.
- Check the row format. Confirm that every non-empty line contains exactly one comma and two valid numeric values in x,y order. Remove headers and units rather than allowing the parser to reject them.
- Enter the title and data. Add the title if you want one, then paste the rows. If the title is intentionally omitted, record that choice in the plot log.
- Generate the plot. Check whether generation succeeded. If no valid point remains, correct the source and generate again because the tool clears an older chart when generation fails.
- Capture the three counts. Write down the valid, invalid, and over-limit numbers exactly as displayed, including zeros. Do not combine the last two categories: invalid rows were checked and rejected, while over-limit rows were deliberately not parsed.
- Record both axis ranges. Read the displayed minimum and maximum on the x and y axes from their labels. Do not estimate them from the picture, and note whether zero falls within either range because the axes do not automatically include it.
- Inspect representative points. Examine several circles in the preview, including points near different limits. Compare their title values with the source rows and note any overlap or repeated coordinates.
- Export the reviewed result. Download the standalone SVG, which represents the same generated output shown in the preview. Record its filename and export date. Editing the title or data clears the generated output and disables the old download, so regenerate and repeat the checks after any change.
Check Scales, Constants, and Overlap
The chart uses independent linear scales. The smallest x value maps to the left boundary and the largest to the right; the smallest y value maps to the bottom and the largest to the top. Intermediate values are positioned in proportion to their numeric distance. Document the displayed ranges rather than describing x positions as evenly spaced, because unequal numeric gaps remain unequal on the chart.
- Constant x axis: When every x value is identical, the x domain expands symmetrically around that value and all points are centered horizontally.
- Constant y axis: The same expansion rule applies independently when every y value is identical.
- Expansion size: The added range is 10 percent of the absolute constant value or one unit, whichever is larger. A single valid point consequently appears centered on both axes rather than causing a division-by-zero error.
- Repeated or close pairs: Markers can hide one another, particularly when identical pairs repeat. Record the overlap and use the source rows and point-title values to distinguish the underlying coordinates.
Include the constant-axis treatment in your notes when it applies. Also state that SVG coordinates are rounded to three decimal places and the chart is not a precision-preserving data format; retain the original rows for exact values, uncertainty, and reproducibility.
Separate the Chart from Statistical Claims
Define the limits of the evidence in the same document. The tool plots supplied pairs on linear axes but does not calculate correlation, regression, a trend line, confidence intervals, error bars, density, clusters, or statistical significance. A visible pattern alone is not a statistical conclusion.
- All parsing, layout, SVG creation, preview preparation, and download preparation happen locally in the current browser; the points and title are not uploaded.
- The SVG is vector output, so lines, text, and circles remain sharp when resized.
- For dense data, categorical color, accessible labeling, logarithmic scales, publication styling, or formal statistical analysis, use a dedicated plotting package and retain the source dataset.
- The chart’s purpose is quick visual inspection. Pair it with appropriate analytical methods when your task requires inference, uncertainty estimates, or formal model checks.
Package a Reusable Plot Log
A reusable log should be concise enough to scan but specific enough to review later. Include these labeled fields:
- Objective: The comparison being made and the meaning of x and y.
- Source: The file, list, or dataset identifier and any preparation performed.
- Exact generation input: The title and complete x,y rows submitted to the tool.
- Summary: The separate valid, invalid, and over-limit counts.
- Scale verification: Both displayed axis ranges and any constant-axis note.
- Point verification: The representative point values inspected and any overlap observed.
- Export: The standalone SVG filename and export date.
- Interpretation limits: Relevant statistical, precision, density, or styling limitations.
A compact note can read: “Generated [title] from [source]; counts: [valid] valid, [invalid] invalid, [over-limit] over-limit; displayed ranges: x [minimum] to [maximum], y [minimum] to [maximum]; inspected [point values]; exported [SVG filename].”
Store that note beside the original data and exported file. If the input changes, create a new version rather than overwriting the old record, then repeat the count, axis, point, and export checks. The goal is to leave a trace more useful than “made graph”: every plotted value, skipped row, visual decision, and delivered file should remain identifiable.
If you're weighing options, How to Document a Line Graph Maker Workflow covers this in detail.