General advice for post-processing ET data

neon

Hi pygazers,

At the risk of perhaps going slightly off topic for this forum, I was wondering what people’s strategies were for post-processing data from eye tracking studies using pygaze / opensesame?

I am sure Edwin’s pygaze analyser is as awesome as pygaze itself, but from my brief look, this appears to focus on producing visual outputs – heatmaps / scanpaths etc. (please do correct me if I am wrong!). I will be looking to determine other common metrics, (time to first fixation, total dwell time etc.) relating to particular, arbitrary areas of interest, and was just wondering what other people’s approaches were.

I’m quite confident I could write a script or program to crudely achieve this by parsing/processing the pygaze log file, but I don’t want to ‘re-invent the wheel’ (or more specifically invent a not-that-great wheel that takes me ages).

Using pygaze/opensesame for gaze-contingent stuff seems great, and pygaze analyser looks fantastic for heatmaps and other visualisations…But, I’m going to be looking for the metrics outlined above for arbitrary areas of interest.
As usual, any advice most gratefully received – even if it is “no choice but to create a script yourself!”

BWs,

Neon

arpit

Hi Neon,
I wanted to know what strategy did you finally take to achieve the post processing of eyetribe data? I am on the same boat as you and wanted some advice before starting up.

neon

Hi Arpit,

Apologies for the delay in replying...
In short, the strategy I finally took is ongoing i.e. I've not had enough time to work on this! However, I'm hoping that will change towards xmas and I'll make some progress.
Essentially I will be writing my own scripts in Python to process the data. In the first instance, these are likely to be quite messy and just be for the purpose of analysing some data that I've got piled up. Afterwards though, I hope to work these in to something a bit more generic, which I will of course happily share.

@Edwin through his PygazeAnalyser work already provides some excellent (and very elegant IMHO) python scripts for detecting saccades, fixations, blinks etc. so a lot of hard work has already been done.
If you are in a hurry and thinking of taking a similar approach to me, I'd suggest looking at Edwins pygaze analyser stuff on github.

Another option you could take would be to use Ogama - I had a look at this, and it appears to be able to do a lot, but I'd just rather spend the effort writing something myself.

Good luck whatever you chose, and let us know how you get on.

Best wishes,

Neon