Research talk:Reading time/Work log/2018-09-26
Add topicWednesday, September 26, 2018
[edit]Today I switched to working on SWAP. The notebook I used today is at /user/nathante/notebooks/investigate time error.ipynb on notebook1004.
DONE Plot Weibull Distributions
[edit]
Reading time measurement by wikipedia language edition and whether the reader is on mobile. Weibull distribution Curves fit to dwell time as measured by the totalLength property, the darker line shows mobile readers. As expected, it appears that the distribution of mobile reading times is lower than that of desktop reading times. The distribution is truncated on the right at 1 hour.
DONE Investigate down the 40 second periodic error
[edit]
Reading time measurement discrepancy by browser. This chart shows a 40 minute oscillating pattern of discrepancies between timestamps of pageLoaded and pageUnloaded events with dwell time as measured by the totalLength property.
DONE (Partially) Compare mobile and desktop distributions
[edit]This may require some understanding of potential issues with mobile. See the above Weibull plot.
Done (Partially) Compare some language editions
[edit]See the above Weibull plot.