Problem: Visualizing longitudinal is challenging when the dataset has large number of panel members. A common ways to visualize such data is to either make a large set of small individual plots such as the example below on the left from Frank 2009, or they resort to 'spaghetti graphs' such as the one on the right

enter image description hereenter image description here

Both ways of visualization have their problems as the charts on the left are hard to read even on A4 due to their size, while chart on the right is even worse as everything overlaps and it becomes completely unintelligible.

Question: Are there any studies/technical reports/academic literature in general that uses some alternative ways to visualize macroeconomic longitudinal macro data?

I am primarily looking for some examples applied to settings like U.S. or EU state level data (i.e. some longitudinal data with $N>20$).

PS: the visualization does not necessary needs to be 'exotic'. For example, I know US can be geographically divided into West, Midwest, South and North and North East. Unfortunately, from my experience even splitting states into 4 smaller panel series produces 'spaghetti' graph (but perhaps there is some other way how to make some trade-off between having one giant spaghetti graph and having individual graph per state?).

PPS: Even though I am looking for examples applied to macro data, I am willing to accept examples from other areas as long as it can be argued that visualization would be useful for macro data as well.

  • 1
    $\begingroup$ there's this thing called a lasagna plot which puts heatmaps of every time series on top of each other. $\endgroup$
    – tdm
    May 9, 2021 at 7:27
  • $\begingroup$ @tdm thanks do you by any chance know of an econ application? $\endgroup$
    – 1muflon1
    May 9, 2021 at 7:35
  • 1
    $\begingroup$ '+1' Tough question, looking forward to some brilliant answers. If it is for online publication dynamic visualizations can be very helpful. If you want to keep the geographic information - assuming the grouping variable related to geolocation - I do not see much of an alternative to doing a video based on a series of mapplots. Or do an entire webaaplication where the user can play around dynamically investigating data properties. Either is offcourse hard to put in a journal :) but then again journals are also evolving often referring to extra resources. $\endgroup$ May 9, 2021 at 8:34
  • $\begingroup$ The king is dead, long live the king! $\endgroup$
    – Giskard
    Jul 23, 2021 at 6:02
  • $\begingroup$ @Giskard thanks (:E $\endgroup$
    – 1muflon1
    Jul 23, 2021 at 8:03


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.