Visualising The Grinder data with R

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Visualising The Grinder data with R

Gary Mulder-3
All,

I've been working on some professional looking visualisations of The Grinder data using the statistical analysis and plotting language R.

The R script I wrote is still very alpha, but is producing some very useful plots already. You can find the R script on Github here (you'll have to change some constants at the start of the script to read your single data log file):


And some sample output plots here:


The histogram plots are an alternative way of showing response times that provides much more detail than simplistic requests over time plots, summary means, and standard deviations. To add context, I've then enhanced the response time histograms by colouring them by response states (e.g. by failed versus successful request, by HTTP response code, or by response length).

Once you get used to R's somewhat unusual syntax, it is very easy to generate ad hoc plots in R and has literally added a whole new dimension to analysing my test results. I'm thinking a stacked bar plot that breaks down overall response time by connect time + time to first byte + rest might be interesting as well.

Feedback and suggestions for added features would be much appreciated! 

Regards,
Gary

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
grinder-use mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/grinder-use
Reply | Threaded
Open this post in threaded view
|

Re: Visualising The Grinder data with R

Joan Picanyol i Puig
* Gary Mulder <[hidden email]> [20160119 20:50]:

> All,
>
> I've been working on some professional looking visualisations of The
> Grinder data using the statistical analysis and plotting language R.
>
> The R script I wrote is still very alpha, but is producing some very useful
> plots already. You can find the R script on Github here (you'll have to
> change some constants at the start of the script to read your single data
> log file):
>
> https://github.com/gjmulder/timeseries-analysis/blob/master/grinder_analysis.R
>
>
> And some sample output plots here:
>
> http://www.perficientur.co.uk/rgrinder/
>
>
> The histogram plots are an alternative way of showing response times that
> provides much more detail than simplistic requests over time plots, summary
> means, and standard deviations. To add context, I've then enhanced the
> response time histograms by colouring them by response states (e.g. by
> failed versus successful request, by HTTP response code, or by response
> length).
>
> Once you get used to R's somewhat unusual syntax, it is very easy to
> generate ad hoc plots in R and has literally added a whole new dimension to
> analysing my test results. I'm thinking a stacked bar plot that breaks down
> overall response time by connect time + time to first byte + rest might be
> interesting as well.
>
> Feedback and suggestions for added features would be much appreciated!
>
Great, thanks for sharing.

I'd probably give a shot at boxplots, also, it'd be nice to have an easy
way to select some tests only.

tks
--
pica

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
grinder-use mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/grinder-use