With Gaia currently in nominal mission mode and sending data to earth, the challenge for the astronomical community is to prepare for the use of what will be at the time of release one of the largest and most complex astronomical catalogues ever produced. Use of parallax data is not straightforward due to the presence of many statistical biases and selection effects. We present an overview of a techniques for correct use of the Gaia parallax information, which relies on statistical modelling of the data in order to infer derived quantities such as distance and absolute magnitude in an unbiased way. The methods rely on a Bayesian methodology and have been applied to case studies on normal stars, variable stars, open clusters and the LMC.