Trend (and Trend Line) in a Chart
TERMThe general direction a series moves over time — and the lines used to show it.
A trend is the general direction a data series moves over time — upward, downward or flat — once you look past the short-term ups and downs. On a chart it is often made explicit with a trend line, a straight or smoothed line fitted through the points to summarise their overall direction and rate of change.
Real data is rarely smooth. Weekly sales, daily temperatures and monthly sign-ups all wobble from one period to the next. A trend is the movement underneath that wobble, and spotting it answers the question most time-based charts exist to address: not "what happened on one day?" but "which way is this heading overall, and how fast?"
Trend, trend line and threshold line: three different things
People often use "trend line" for any extra line drawn on a chart, but the terms are not interchangeable. If you are looking for the word for a line that marks a limit or target, the term you want is a reference line (also called a threshold line, target line or benchmark line) — not a trend line.
| Term | What it is | Where it comes from | Example |
|---|---|---|---|
| Trend | The overall direction of the data | A pattern you observe in the data | "Sign-ups are rising" |
| Trend line (trendline) | A line fitted through the data to summarise that direction | Calculated from the data, e.g. by least-squares regression or a moving average | A straight line sloping up through monthly sign-ups |
| Reference / threshold line | A fixed line marking a limit, goal or benchmark | Chosen by the chart maker, not calculated from the plotted points | A dashed line at a sales target or a safety limit |
| Baseline | The starting value the data is measured from | Usually zero, or a reference period | The zero line bars grow from |
Threshold lines are usually horizontal (a constant value) and often dashed or labelled, such as "Target: 20" or "Limit". Some fields have their own names: control charts in quality management use control limits, and finance charts use support and resistance lines. A trend line, by contrast, moves with the data — if the points change, the trend line changes too.
A worked example
Suppose a newsletter gains these numbers of new subscribers over eight months: 10, 14, 11, 16, 15, 19, 17, 22. Month to month the figure bounces — it falls in months 3, 5 and 7 — so looking at any single step could suggest decline. Looking at all eight points together shows a clear upward trend.
Fitting a straight trend line by least squares gives roughly y = 8.96 + 1.45 × month. The slope, about 1.45, is the useful number: on average, monthly sign-ups grew by around one and a half per month over the period. The team's target of 20 per month is the threshold line; the data only crosses it in month 8, and the trend line reaches it around month 8 as well, which is a more honest summary than celebrating one strong month.
Trend vs noise vs seasonality
Most time series can be thought of as three parts layered on top of each other:
- Trend — the long-run direction (growth, decline, or flat).
- Seasonality — a pattern that repeats on a fixed calendar cycle, such as higher retail sales every December or lower electricity use every spring.
- Noise — irregular, unpredictable variation left over once trend and seasonality are accounted for.
A December spike is not a trend if it happens every December; compare each month with the same month in previous years to see whether the underlying level is genuinely changing. Likewise, a single extreme point is an outlier, not a trend.
How to identify a trend on a chart
- Check the time span. Read the x-axis to see how many periods are shown. A handful of points can make random movement look like a direction.
- Check the scale. A y-axis that starts far above zero, or a compressed time axis, can make a gentle slope look steep. See misleading charts for examples.
- Compare the start and end, then the middle. If later values are consistently higher than earlier ones, there is an upward trend even with dips along the way.
- Smooth it. Add a trend line or a moving average (each point replaced by the average of the last few periods) to see the direction without the noise.
- Describe it with a rate. "Up about 1.5 per month" is more useful than "going up".
Types of trend line
Spreadsheet tools such as Excel and Google Sheets offer several trendline types. Choose the one that matches the shape of the data, not the one that looks most impressive:
- Linear — a straight line; fits data that rises or falls by a roughly constant amount each period. This is the same idea as a regression line on a scatter plot.
- Exponential — fits data that grows by a roughly constant percentage each period, such as compound growth. Often clearer on a log scale.
- Moving average — not a formula but a rolling average; good for removing short-term noise without assuming a shape.
- Polynomial — a curve that can bend; fits rise-then-fall patterns but easily overfits, so use it cautiously.
Reading a lasting trend from a short window; extending a trend line far beyond the data (past movement does not guarantee future movement); calling a seasonal swing a trend; and assuming a trend shows cause — two series trending together are correlated, which is not the same as one causing the other.
Where trends show up
Trends are easiest to see on a line chart, where connecting points draws the eye along the overall slope; an area chart works similarly. On a scatter plot the same idea appears as the direction of the point cloud, summarised with a regression line. You can plot your own series and judge the trend with the line chart maker.
Frequently asked questions
What is the line on a chart representing a threshold called?
It is usually called a reference line, threshold line, target line or benchmark line. It is a fixed value chosen by the chart maker, such as a goal or limit, and is different from a trend line, which is calculated from the plotted data.
What is the difference between a trend and a trend line?
A trend is the general direction of the data over time. A trend line is a line fitted through the data points, for example by least-squares regression or a moving average, that makes that direction visible and lets you state its rate of change.
How do you tell a trend from random noise?
Look at a long enough time span, compare early values with later ones, and add a trend line or moving average. If the smoothed line keeps moving in one direction while individual points bounce around it, the movement is a trend rather than noise.
Is a seasonal pattern a trend?
No. Seasonality repeats on a fixed cycle, such as every December. A trend is a longer-run change in the underlying level. Comparing each period with the same period in earlier years separates the two.
Can a trend line predict the future?
Only cautiously. A trend line summarises past data, and extending it forward assumes conditions stay the same. The further you extend it beyond the data, the less reliable the projection becomes.