I propose to temporally visualize uncertainty
in financial data modelled with Brownian motion.

In particular, with appropriate estimates of parameters, it is possible to
construct a distribution over Brownian bridges connecting each pair of data
points in a time series. I propose to repeatedly sample from this
distribution in time, displaying each sample to the user as possible
"imputed" data. That is, given a time series of "known" data points, at
each time, the user will see a new Brownian bridge connecting each pair of
known data points, sampled from the modelled distribution. Combining this
approach with a decay in opacity for each newly sampled Brownian bridge,
the user should get a very intuitive sense of the underlying modelled
uncertainties of the time series. Moreover, the approach should be amenable
to vectorized computation on GPU hardware.