The tool below will give you an idea of where the most relative demand is for Uber rides during a particular hour based on the parameters you select. It is based on 66 million Uber pickups from January 2016 through December 2016. The tool only deals with demand for rides, but does not provide any insight into the supply of drivers. So keep that in mind if you're a driver using the tool. If you are looking for more interesting insights from the data, go here. You can find out more about the process behind the tool below.
The data is sourced from a Freedom of Information Act request that FiveThirtyEight filed with the NYC Taxi & Limousine Commission. The background and data can be found here. The tool was also inspired by an awesome post by Todd. W. Schneider.
Subsequent to that data release, the NYC TLC began releasing monthly data for what they call "For Hire Vehicles" which includes Uber. As of March 2017, the latest available data was for the month of December 2016. If you want to explore the data yourself, it can be found here. It includes the data disclosed to FiveThirtyEight previously mentioned.
Why aren't other variables included? Well, I experimented with a few other potential predictors. I added variables to account for whether it was a holiday and whether the next day was a holiday, with the hypothesis being that both of those factors would influence rider behavior (if tomorrow's a holiday, maybe riders are more likely to stay out late the night before, etc...). And including those did indeed improve the quality of the predictions. However, their inclusion caused the amount of predictions to increase substantially without offering a significant improvement accuracy. So I decided to leave them out to simplify the presentation and usefulness of the tool above.