21 Jun
2019
21 Jun
'19
1:17 a.m.
Hi all,
I was wondering if the community had any suggestions of toolboxes to plot the output of
simulations (ideally in python) and make publication-quality figures.
I am mostly thinking of eg spike rasterplots or heatmaps (time x variable) with large
datasets. Ideally in such a way that those can be aligned with other figures easily (spike
histograms, LFP, you name it)
This is a very general problem that I suppose a lot of us have been confronted with, and I
actually haven't found so far a good solution:
-matplotlib (python) is robust and can handle large datasets but looks bad and is a
nightmare to deal with when trying to make figures with multiple subplots of specific
sizes etc
- altair (python) looked promising and versatile but is actually struggling with large
datasets
- The existing projects for (analysis and) visualization of neural data that I have
checked (elephant spykes etc) are either quite specific/limited or not very active.
Additionally they are focused on real ePhys data: either multiunit spike data or single
neuron continuous data, while simulations provide both.
Happy to get any suggestions and share some code !
Thanks, Tom
21 Jun
21 Jun
10:22 a.m.
Dear Tom,
I’m not aware of any good neuroscience-specific tool for this.
“matplotlib…looks bad” - this is true if using the default settings; however, it can be
tweaked to give very nice output (although this is often very painful to do). Are you
aware of Seaborn (https://seaborn.pydata.org)? This is a package built on top of
matplotlib which gives nicer output by default.
You might also want to have a look at https://holoviews.org. I haven’t used it myself, but
the underlying concept is interesting.
Concerning a plotting package specifically for neuroscience (including simulation data),
there is a plan to develop a companion to the Neo and Elephant packages, which would
provide plotting tools both for raw data (e.g. Neo AnalogSignals, SpikeTrains) and for the
results of data analysis produced by Elephant (heatmaps, spike histograms,
cross-correlation plots, etc). Michael Denker could give you more information on the
current planning, but in any case the idea is to have an open development process on
Github so anyone can contribute. The NeuralEnsemble mailing list
(https://groups.google.com/forum/#!forum/neuralensemble) would be a good place to discuss
this.
Regards,
Andrew
On 21 Jun 2019, at 01:17, TOM BUGNON
<bugnon(a)wisc.edu> wrote:
Hi all,
I was wondering if the community had any suggestions of toolboxes to plot the output of
simulations (ideally in python) and make publication-quality figures.
I am mostly thinking of eg spike rasterplots or heatmaps (time x variable) with large
datasets. Ideally in such a way that those can be aligned with other figures easily (spike
histograms, LFP, you name it)
This is a very general problem that I suppose a lot of us have been confronted with, and
I actually haven't found so far a good solution:
-matplotlib (python) is robust and can handle large datasets but looks bad and is a
nightmare to deal with when trying to make figures with multiple subplots of specific
sizes etc
- altair (python) looked promising and versatile but is actually struggling with large
datasets
- The existing projects for (analysis and) visualization of neural data that I have
checked (elephant spykes etc) are either quite specific/limited or not very active.
Additionally they are focused on real ePhys data: either multiunit spike data or single
neuron continuous data, while simulations provide both.
Happy to get any suggestions and share some code !
Thanks, Tom
_______________________________________________
NEST Users mailing list -- users(a)nest-simulator.org
<mailto:users@nest-simulator.org>
To unsubscribe send an email to users-leave(a)nest-simulator.org
<mailto:users-leave@nest-simulator.org>
--
Andrew Davison, PhD
Department for Integrative and Computational Neuroscience (ICN)
Paris-Saclay Institute of Neuroscience
UMR 9197 CNRS/Université Paris Sud
1, avenue de la Terrasse
91198 Gif sur Yvette
France
Tel: +33 1 69 82 34 51
http://www.andrewdavison.info/
1764
days inactive
1765
days old
1 comments
2 participants
participants (2)
-
Andrew Davison -
TOM BUGNON