Previous proposalScalability truths and serverless architectures: why it is harder with stateful, data-driven systems
Next proposalMath for data science
Atlas: GO-JEK’s real-time geospatial visualization platform
Submitted by Ravi Suhag (@ravisuhag) on Tuesday, 20 February 2018
Section: Full talk Technical level: Intermediate Status: Confirmed & Scheduled
We have billions of GPS points flowing through our data pipelines daily in real-time and drive decisions like driver allocation, surge pricing, driver incentives and more. This poses intriguing challenges in finding actionable insights from spatial data in real-time.
At GoJek we built Atlas in an attempt to make it easy for teams within GO-JEK to visually explore this flood of geospatial data. For example:
- Our City Ops teams need to have near real-time information of the current supply and demand distribution.
- Our Data Science team needs rich exploratory interfaces to tackle multidimensional data for experimental analysis.
In this talk I will discuss about the need, challenges and architecture of ATLAS.
- Brief about Speaker and GoJek
- State of data at GoJek
- Challenges in making realtime decisions with data
- Atlas Introduction
- Data pipeline architecture
- Atlas Architecture
- Atlas metric streaming
- Atlas dimension mapping
- Atlas data experience
- Road Ahead
Ravi Suhag works on the data enginerring team at GoJek which was responsible for building ATLAS end to end.
To know more about the speaker please visit: http://www.ravisuhag.com