Follow Datanami:
August 1, 2012

What it Takes to Deliver Real-Time Traffic Info

Datanami Staff

Providing real-time traffic updates is a practical and feasible application of the big data analysis technology that exists today.

According to real-time streaming big data platform vendor, SQLstream, which aims to integrate and quickly analyze live data feeds,  “High-performance systems such as SQLstream Transport are able to transform large volumes of raw GPS data into real-time actionable information.” The transformation in question took place on the roads of Venezuela in the form of SQLstream’s ETL Connector for Google Big Query.

At first glance, using GPS data to monitor traffic flow does not seem too difficult. After all, local newscasters have been giving traffic updates by sight from helicopters for awhile. However, a GPS submits plenty of data, including the important location, speed, and destination variables. Further, there may be tens of thousands of cars on the roadways around Caracas, and if half of them are reporting GPS data, that ends up being a lot of data to sift through for the capital alone.

The architecture works as such: GPS collectors feed information into an interpolator. SQLstream is collecting information for each ten-meter strip of Venezuelan roadway and sometimes individual strips are unoccupied by a GPS. The interpolator fills in the rest and sends its information to both Google Big Query and to traffic flow, congestion detection, and condition prediction algorithms.

The final result is a map of Venezuela with roads color-coded based on the average vehicular speed relative to that of the posted speed limit and with pink push pins denoting a sudden change in the average speed for that particular ten-meter stretch (the example in the video is of an area whose average vehicular speed drops from 60 to 25 miles per hour in only fifteen minutes). The map is elegant and informative.

What is interesting to note is the choice of location for this case study. Perhaps it was advantageous to select a locale whose traffic patterns were not so well-documented (if, for example, SQLstream had indicated that it is difficult to get around Washington D.C.’s beltway at, well, pretty much all times, it would have been easy to sigh and move on).

 It is more likely, however, that testing in Venezuela would offer significantly less data. While the GPS becomes more affordable, the average Venezuelan has significantly less disposable than the average American. That equals fewer GPS systems, which equals less data. If there exists less data, the goal of providing a comprehensive traffic map becomes more reasonable. However, this is simply speculation.

What is not speculation is that SQLstream, partnered with Google, seems to have provided an impressive traffic map of an entire South American country.

Related Stories

Six Super-Scale Hadoop Deployments

Another Brick in the Hadoop Wall

Partnership Targets BI Scalability