Machine Learning and Graph Analytics on GPU-Accelerated Data Science

During the first part of our series on Seattle Parking dataset we explored the functionality of cuDF, dask_cuDF and BlazingSQL. However, the distance we calculated from the Space Needle to the closest parking spots assumed we traveled in straight lines which is highly improbable in an urban setting. Moreover, the frequentist approach we took to calculate the parking occupancy is a very coarse estimate of what to expect in terms of the actual parking occupancy.

In this webinar we alleviate both of these issues. First, we build a regression model to predict the actual occupancy at each parking spot given the time of the day and a day of the week. Second, we use a graph dataset of King County roads, kindly prepared and donated by John Murray from Fusion Data Science, to properly calculate the walking distance as well as the actual route to each parking spot.

[guowuzhang]

Thanks a lot for sharing you knowledge. What is the password if i want to use sudo?