The time series is one of the most basic and intuitive visualizations.
A time series graph allows the viewer to focus on one single metric while exploring that metric’s behavior over time. There are many advantages to a regular time series that solidify its popularity. The simplicity and ubiquitousness of the time series graph means that it rarely needs to be explained. It also allows for easy comparisons of slopes that indicate the rate of change.
While the time series chart has many advantages, under some circumstances, it does not correspond well to the human perception of time. Humans conceptualize time in perspective, starting from the present and thinking into the past (or the future). Events that happened earlier in the day are very vivid, more so than events that happened last week, and even more so than events that happened last month. If you examine the past year of, say, unique visitors to your website, you might want more resolution to make more detailed comparisons dedicated to the past day than the past week than the past month, etc.
One option is use a regular logarithmic scale to represent time. This addresses the perspective issue with visualizing time, giving a more fine-grained look at more recent periods of time. However, logarithmic scales are unfamiliar to people and would require an extended explanation, making it an impractical option.
A better solution is to represent the time axis as a piecewise linear time scale (PLTS) that gives equal resolution to increasingly bigger segments that make sense on the human cognitive level (e.g. day, week, month, year, etc.).
This visualization attempts to keep the detail proportional to the “perceived importance” of the time in question. It would require more explanation than a regular time series chart but it would also provide some novelty and attract the eye. The added time perspective comes at a price. It becomes harder to compare slopes of lines between different scale ranges. This visualization would also only work if one end of it is anchored in the present and therefore would not be suitable for displaying time data from the past or support the regular click and drag to shift time interaction technique usually afforded by (interactive) time charts.
Some of the problems outlined above can be remedied by allowing the PLTS plot to seamlessly emerge from a regular time series plot by a simple animation. The animation would allow the viewer to intuitively understand how time is being transformed because they see where the parts of the time series plot are mapped to on the PLTS plot.
The PLTS is not a general replacement for the time series plot, but if the time range of the examined data is anchored in the present, it could be a way to display a large segment of time without sacrificing detail where it really matters.