A good data vis tells a story. Good, effective stories are clear of confusing distraction, tightly focused on the core message.
- The plotted lines should be the most prominent part of the graph. Use medium stroke, keep gridlines faint, choose meaningful colors, aim for 6 lines or less, vary line style sparingly. Use color or annotation to draw attention to the key line / event you want the reader to pay attention to.
- Keep the data unambiguous. Decrease chances of misinterpretation by using uniform time spacing, labeling axes directly instead of in the title, place labels near the lines instead of in a legend, and avoid dual y-axes. Because viewers read slope more than absolute position, ensure the aspect ratio doesn't artificially flatten or steepen trends.
- Provide clear context. Use a concise and informative title that states the relationship or trend, visually show the data range represented in your plot, and annotate important events directly on the plot.
- Be ethically transparent. Break the line or use a different style for imputed segments. Explicitly disclose if you are smoothing your data and what method you used.