Line chart with far too many lines. They become illegible.
A line chart with a huge key that squashes the chart.
A simple line chart with three lines, telling a simple story about the rise of meat consumption in China and Brazil
A line chart showing the rise of chicken as the world's favourite meat
A line chart with not enough information - sometimes four lines hides the bigger picture
A line chart that doesn't have any data to be interesting
A line chart where all of the lines are visible but greyed out. The key lines for the story are pulled out in a highlight colour.
The whole dataset is visible but greyed out. Four key lines are highlighted.
Small multiples - the same chart repeated, with each country getting its own individual chart
Small multiples - this time with area charts, rather than line charts
A slope chart can often be a better choice when you have more than five lines to display
Slope charts can often take more lines than a line chart, because they only show two points in time
A dot chart showing decline in fertility over time
A dumbbell chart showing decline in fertility. Both the start and end points are visible.
A default line chart in Powerpoint
A default line chart in Flourish - all shades of one colour
Three lines on a line chart - colours clearly differentiated
Lines are about similar things - different types of sea creature - so colours are similar
Using a bright colour to pull out the line that's going to be of most interest to your audience. Here the UK's poor showing in the number of books published compared to other western countries
Using a highlight colour to show the most interesting story. One of these lines - Romania's - is behaving differently to the others
Shades of a single colour sometimes work as an alternative to lots of clashing colours
Separating out the lines into groups and using the same colour for highlighted lines can also work.
Small multiples can sometimes be clearer than a busy line chart
Small multiples with an average line can be an effective way of showing where each country sits
Nothing to see here
Colour can help signal a hierarchy of importance
Using colour to signal what the line is doing - rising, dropping or flatlining - can also work
India grows the most bananas. The bright yellow-orange fill tells this story effectively.
The same yellow-orange struggles as a line
Sometimes you need to be less subtle with line colours - e.g. with a bright red or similar primary colour
A dark background can also help bright colours to punch out
A thicker line width is sometimes needed to let bright colours shine
Don't do this. Or anything remotely resembling this.
Sometimes the story only makes sense with ghastly colours
No colour (black) is sometimes the best highlight colour of all
A complex line chart with lots of multi-coloured zig-zagging lines
A line chart showing lines that overlap and with too many datapoints
Showing every data point can lead to a story of wild fluctuation, and the bigger story can be lost
In this CO2 concentration chart, is it better to show the general trend than the seasonal variation?
Plotting every datapoint can just leave you with a blurry mess
Using one datapoint per hour or per day can create clearer charts
Using straight lines between points means key moments and changes are clearer
A curved line means that you often undershoot or overshoot each datapoint, which can undermine the drama when the line changes direction
The straight lines make the changes in the data obvious
Smooth lines give us just one peak per rush hour - which isn't true
If you have Census data, once every 10 years, then of course show all the data on each line
When you have one datapoint per year, you can usually show all the data
Vertical bar charts - starting correctly at zero - showing the number of insect parts you're legally allowed to have in US food. (Thousands, basically).
Bar chart that doesn't start at zero. Now the bars are deceptive.
A bubble chart showing that in the US you are allowed 149 dead insect parts in 100 grams of flour, and 59 parts in 100 grams broccoli.
A bubble chart with the bottom third of each circle chopped off. Ridiculous and deceptive.
By starting the line chart above zero, it looks like a lot has changed, when it hasn't.
By starting the line chart at zero, the story is retained: there have always been too many privileged people in Parliament (around a third of the total MPs) - and there still are.
By not starting at zero, the meaningful range of data is shown. Here it is the speed of cyclists in London. We can start at 11mph, because the average is never lower than 12mph.
If we deliberately start at zero, then the important changes in cycling speed at rush hour peaks are flattened out.
Sometimes you start at zero, even if you don't technically have to (it's not in your data). Just because zero is a useful baseline or starting point for the dataset.
Not starting at zero here seems pointless. It undermines the story of evolution from things moving slowly, to things moving more quickly.
For a height chart, you might start at zero.
Chart showing how life expectancy has changed during Russian history, falling to 16 during the Second World War
You would definitely not start at zero here. This chart shows how life expectancy dropped by a year in key western countries, and that drop is important, and would be invisible if you started at zero.
By making the UK line intersect with the x-axis, it looks like life expectancy has fallen to zero.
If you start above zero, no fill.
If your y-axis starts above zero, you should never fill the area below the line.
A chart showing John Travolta's filmography over time, from the heights of Grease and Pulp Fiction, to the depths of Battlefield Earth.
A chart showing how 13 out of 33 of Spielberg's films have received 5 stars from Empire magazine, and only one (Hook) has received less than 3.
Chart showing the global decline in fertility rates - each datapoint has a circular marker, making it hard to distinguish the lines
Population chart from 0 AD to 2021 - with a marker for each datapoint. For the first 1,700 years, the datapoints are every few years. From 1700, the datapoints are annual and all smudge together.
A chart showing how the Lib Dem and Labour votes mirror each other. Each election is shown with a clear marker.
Chart showing contraceptive use in Japan and Lesotho. Lesotho's reporting is intermittent and this is shown by the markers.
A chart showing the decline in the belief that having children before marriage is wrong, from 70% in the 1980s to 42% in 2012. A large circular marker is shown for each datapoint.
A slopechart showing how we have become more tolerant of homosexuality, sex before marriage, and having children before marriage. However disapproval of extramarital sex remains high.
A chart showing the decline in Irish population after the Great Famine. Markers indicate the highpoint before the Famine, and the low point before Ireland joined the EU and population started to rise.
A chart showing how people no longer care about the gender of their boss. There is a marker for the final datapoint of each line.
Chart showing Facebook share price falling. No data labels makes it hard to work out the exact degree of change.
Chart showing Facebook's shares dropping. Data markers make it clear what the starting value was - and then its peak and lowest value.
A succession of charts showing that the more lines you have on your line chart, the more data markers should be dropped
A series of charts showing bad data markers - triangles, diamonds, photos and lots of floating numbers
Chart showing McDonalds share price recovering after the Pandemic, but the lines are too thick for the small changes to be visible
Fertility rates chart where the lines are far too thick and the individual lines are too hard to untangle.
As the lines become thinner, all of them become more readable. The fall in fertility rates across the Middle East is now more visible.
All the lines are thick and the story of the UK's decline after Brexit is not clear.
The main story line is now thicker than the others, so the UK's trading weakness after Brexit is now clearly visible.
Use a grey or dotted line to show an average. Here we see the average inflation rate across the EU and how the Baltic States have much higher price rises - due to the Ukraine invasion
Use a grey or dotted line for averages and projections. Here we can see how Kuwait's defence spending has always been higher than the global average, and shot up to over 100% of GDP during the Gulf Wa
Line chart showing McDonalds' share price recovering after Covid. Moderna and Zoom's value soars during Covid-19 then plummets below McDonalds. The lines are thinner.
Chart showing how men have increased in height in the last 100 years, especially in the Netherlands. Only in a few sub-Saharan countries have average heights declined.
Chart showing the catastrophic decline in Arctic ice volume, especially in recent years, when it appears to be accelerating
A line chart showing how Brazil and China's meat consumption has increased. There is a key above the chart that makes it hard to quickly see what each line means.
A chart showing how meat consumption in Brazil and China has increased. Now the lines are directly labelled so you don't have to keep comparing the lines to a separate key.
The same line chart as above - showing how chicken is the world's most popular meat - but without the angled connecting lines
The same line chart, but with some of the lines deleted, so the story of chicken's popularity as a foodstuff can be more clearly seen
A line chart showing how there are more boys than girls born in most South East Asian countries. All the lines are bright colours and it hard to see the story.
The same line chart, but here only the key outliers are highlighted - China, Hong Kong, South Korea (which have or had a lot more male children born) and Mongolia (with more girls born).
A line chart with grid lines, tick marks and annotation lines obscuring the data lines
A line chart with gridlines removed, and only essential axis lines present
A line chart showing Facebook's stock price dropping after the anti-trust lawsuit against it. Only the data line and a y-axis line is necessary.
A line showing how average heights have grown in South Korea but plateaued in sub-Saharan Africa. Gridlines are a useful reference point here.
Where zero is present in the dataset, particularly if you have negative numbers, you need to have a clear zero line.
Where you don't have zero in your dataset, or if it's not meaningful, you don't need an x-axis, just the labels. This chart shows how US life expectancy plunged after Covid - more than its peers.
 A chart showing that sadness levels in the UK rose during the Covid lockdown, when Russia invaded Ukraine and when Liz Truss (briefly) became Prime Minister. Annotation lines are subtle.
Life expectancy in China dropped by 15 years during Mao's 'Great Leap Forward' between 1958-62. This is annotated by a grey box overlaid on the chart.
A line chart showing how the abundance wildlife has declined by almost 80% since the 1970s. Grey lines are used to show the upper and lower confidence bars.
A line chart showing the decline in the abundance of wildlife since the 70s - a drop of almost 80%. The upper and lower estimates are shown by a shaded grey area.
A line chart showing the decline of wildlife in Africa and South America. Error bars on the lines make both of them virtually unreadable.
Two lines charts showing the decline in wildlife in Africa and Latin America. They are side by side and a grey shaded area shows the upper and lower estimate.
A bar chart showing the Top 10 baby boy names in England and Wales. Oliver is first. Some of the bars are unlabelled so it is impossible to know the names.
A pie chart showing the number of lines each character has in Hamlet. Hamlet has around a third of the total. Most of the other slices are unlabelled so it is impossible to understand.
A chart showing Facebook's declining share price after the anti-trust case brought against it. Only two labels - the start and end points - are required on the x-axis.
A line chart showing bond yields - and how they've declined in value during the spring and summer of 2019. Only two x-axis labels here is confusing.
A line chart showing the peaks (Blow Out, Pulp Fiction) and troughs (Battleship Earth) in John Travolta's career. The six x-axis labels are evenly spaced.
A line chart showing how average male height in South Korea has consistently climbed, while height in sub-Saharan Africa has dropped. The start and end value are on the x-axis, then 20-year intervals
A line chart with vertical x-axis labels. Unreadable.
A line chart showing the high percentage of foreign-born populations in Gulf states. In Qatar it is almost 90%. The global average is 3%. X-axis labels are horizontal and readable.
A line chart showing the decline in civil service numbers in the UK since the War. The line is inset, there is a gap before the start for no obvious reason.
A line chart showing the decline in civil service numbers in the UK since the War. The line goes edge-to-edge and is therefore less confusing.
A 2D chart. The lines can be easily decoded.
A 3D line chart showing the same data. The direction and orientation of the lines is now hard to decode.
A Powerpoint chart showing a single line changing over time. The datapoints are easy to decode.
A 3D line is confusing, as the peaks and troughs cannot be easily compared.
prev / next