1

I would like to use a line chart but I cannot decide if I should add markers to the lines or not.

  • On one hand adding markers seems like extra noise
  • On the other hand they make location of the data points more clear. This might be handy if the slope of the line doesn't change, but is this really a valid point?

What is the purpose of placing markers?

Line chart with and without markers

  • It is also possible that a line chart isn't the best way to represent the information you want to display... generally it is used to show trend over time but since you have multiple data series it makes the trend information less of a focus. Line charts are most powerful when one or two data series is plotted since it is hard for people to process more complex information. – Michael Lai Apr 24 '17 at 22:29
  • @MichaelLai This specific chart is about the effect of the trend of the blue line on the trend of the yellow and green line. – totymedli Apr 25 '17 at 0:30
  • 1
    I know the images you showed may not be representative of the actual data, but if you want to show the effect of the change of value for the blue data set on the yellow and green data sets it is also possible to use the blue value as the baseline and just plot the change in the yellow and green data sets as a percentage of the change in the blue value to reduce the number of data sets displayed on the chart, or other types of normalization depending on the relationship between the three data sets. – Michael Lai Apr 26 '17 at 11:59
  • @MichaelLai Great idea! I choose this because we are interested in the values/trend of the three data set and the effect too. Also, this might be easier to reason about for the users, but your idea would make up a great second (additional) chart! Thanks! – totymedli Apr 27 '17 at 18:12
3

While it's not always the case, the most common reason is that markers represents the data that has been entered/registered, while the lines represent the trend or "traveling" within an X period of time where data goes from N to N', N'' and so on.

Edit: see example below: Feb had 20, yet it's impossible to tell if there was data for Feb or if the line passed thru Feb by going from ene to mar. Or in other words: We want to visualize data, yet we can't give proper information and leave data as ambiguous (which is the exact opposite of what we should do)

enter image description here Image Source: Uxpanol.com

When you have peeks and valleys, we can assume the vertex is a data point. However, let's assume data goes always in more or less the same direction. For example, it goes always up. Without the data points, you won't be able to visualize whether the specific axis has data or not. Just imagine we have day1, day2, day3, day4 on the X axis. Then 100, 200, 300, 400 on the Y axis. Without data points, you could have 2 scenarios. We have data for all 4 days, or 3 days, or just day1 and day 2 going from 100 to 400, with day 2 and 3 registering no data. You can see this woudn't make any sense

And that's why you add data points: for increased visualization.

  • I agree, but what if I always have data for every month? Especially if this is expected/assumed by the users. – totymedli Apr 25 '17 at 0:26
  • It will confirm data is there, without any ambiguity. Either way, visualization will be improved – Devin Apr 25 '17 at 14:32
2

It depends from the use cases you expect this graph to be used. If it is very importand for users to know exactly which are the data points, then you must add them. But the most typical use case for graphs does not require data points to be explicitly displayed, and in this case you can remove them to increase data-ink-ratio.

Additionaly consider removing the grid lines.

enter image description here

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.