Start with the message, surface and viewing conditions. A palette for a printed chart has a different job from paint in a room. This guide does not assign a fixed psychological effect to blue, red or any other hue, and it does not promise a productivity gain.
Two grey labels that do not have the same contrast
For solid sRGB text on white, the supplied calculation gives approximately 4.48:1 for#777777 and7.00:1 for#595959. Both might be described casually as “grey,” but the first is below the4.5:1 threshold for ordinary text in WCAG's contrast criterion; do not round4.48 up to a pass. Black on white gives 21:1 and provides a useful check on the calculation.
The source converts each sRGB channel to linear light, calculates relative luminance and uses(Llighter+0.05)/(Ldarker+0.05). These are constructed solid-colour pairs. A photograph, transparency, gradient or hover state needs its actual background checked separately. Meeting this one contrast condition does not certify a whole page or remove the need for non-colour labels.
The input and output for color-psychology-explained are in the calculation results. Run the standard-library Python source to reproduce them; it does not read your files or send any data.
Define what the colour is doing
List whether a colour identifies a category, attracts attention, separates a layer or simply decorates. For each meaningful use, write the information that would be lost if the colour disappeared. This reveals which parts also need words, shapes, line styles or position.
The W3C explanation of use of colour says colour must not be the only visual means of conveying information. This is an accessibility principle, not a theory that one colour changes everyone's mood.
| Element | Intended meaning | Additional cue | Check in the actual output |
|---|---|---|---|
| Fictional chart: red series | Planned spending | Label “Plan” and dashed line | Read it without naming the colour |
| Fictional chart: green series | Recorded spending | Label “Actual” and solid line | Check greyscale and normal view |
| Your element | Meaning or decoration | Label, shape or other cue | Screen, print or material sample |
Check readability separately from preference
Inspect text against its actual background, including links, small labels and projected or printed output. A palette can be attractive while making a label difficult to read. For web work, use the relevant WCAG contrast guidance and a suitable checker; this article does not certify a palette from colour names alone.
For a room or physical object, try a small sample in the actual light before a large change. A screen swatch is not the same material or illumination. Describe what you observe rather than assuming a warm/cool chart predicts another person's comfort.
Compare alternatives with the same task
Fictional comparison: two versions of a schedule use different accent colours but identical wording. Ask whether an unfamiliar reader can locate the deadline and distinguish confirmed from tentative items. If neither version communicates that distinction, change the labels before debating the accent. The scenario is a proposed check, not a measured user study.
Finish when the message remains understandable without relying on hue alone and the intended output is readable. Preference can then guide decorative choices within those constraints.