My dream project would be to do something like this but with better data input, i.e. with gps loggers that send more pricise positions more often, maybe linked with vehicle-mounted cameras to better understand the actual causes of delay in a more qualitative way.

Show Post

So reading the definition of a clean run, I can foresee that busy daytime stops will show up as delayed when I think what we're interested in teasing out is signal or traffic delay.

Show Post

Put it more simply, we suffer from a quantity over quality syndrome in research since collecting and crunching big data became much easier.

Show Post

My main argument is that "a bunch of guys with a chronometer and a note pad spending a few entire days going up and down a line while taking note of the apparent causes of delay" might be more effective than a dataset whose main strengths is just having a lot of datapoint.

Show Post

Gotta love when he suggests cyclists take the subways instead. Of course he won't suggest that option for drivers.😡

Show Post