Uber’s latest predictive model will deduce where you’re headed and it’ll be accurate 74% of the time.

With the use of data analytics, the data team at Uber incorporated classic Bayesian statistics to set up a model that infers “where people ultimately want to go, rather than where people may specify where they want to be dropped off, in order to get there,” explains Uber’s Ren Lu in a blogpost.

The blog post explains the basic method of the research:

We took the riding patterns of over 3000 unique riders in San Francisco earlier in 2014 (anonymizing the data to protect privacy.) Each of these trips had been “tagged” by the rider: when requesting an Uber, the rider had filled in the destination field. We assumed that this represented the true destination the rider wanted to go, creating a gold standard against which we can compare the predictions of our model.

Amongst the comprehensive stats, the post enunciates that the Uber model works around three crucial variables, “priors”, which are the rider’s previous destinations, the discernable patterns of all riders combined and the popular places frequented by average riders.

“Extensions of this project involve building more complex priors and likelihoods,” the blog said. Uber believes that this model is a step towards making the “Uber experience” better.

Based in San Francisco, and serving cities worldwide, the outcome of implementation of this model remains to be seen.

Read more here.

(Featured Image: Uber)

Previous post

The Week in Big Data-- August 4, 2014

Next post

Google Reveals Revolutionary Datawarehouse Mesa