Presented By O'Reilly and Cloudera
Make Data Work
September 26–27, 2016: Training
September 27–29, 2016: Tutorials & Conference
New York, NY

Delivering near real-time mobility insights at Swisscom

Francois Garillot (Swisscom)
4:35pm–5:15pm Wednesday, 09/28/2016
Spark & beyond
Location: Hall 1B Level: Intermediate
Average rating: ***..
(3.75, 4 ratings)

Prerequisite knowledge

  • A familiarity with the basic tenets of Spark (batches, DStreams, RDDs)
  • Basic understanding of real-time analytics and event-driven platforms
  • What you'll learn

  • Know how key choices drive a technical stack's composition
  • Understand how to connect big data tools such as Spark, customer focus, and the values embodied in a product
  • Description

    Swisscom, the leading mobile service provider in Switzerland, also provides data-driven intelligence through the analysis of the data created by its mobile network. Its Mobility Insights team works to help civil administrators, tourism and marketing professionals, and many others understand the flow of people through their locations of interest. François Garillot outlines the platform, tooling, and choices that help achieve this service and some challenges the team has faced, before exploring in depth the task of understanding the speeds of populations through a path of interest.

    François offers an overview of the design of Swisscom’s big data infrastructure, which features Scala, Kafka, and Spark as pivotal tools, focusing on the multiple components that allow unified and reliable access to the high-throughput data flowing from the telecommunication network and lead to a single platform that lets engineers answer heterogeneous questions quickly. Along the way, François explains the technical challenges of moving into real-time analysis and fast data, a key feature of the speed measurement task.

    François also proposes possible solutions to technical challenges, such as selecting interesting datapoints out of a millions coming in every second, sessionizing when no batch interval seems to make clear sense, and the importance of checking ground truth, and explains how privacy protection—crucial to Swisscom and the Mobility Insights team—is constitutive of both the data filtering and the questions the team chooses to tackle.

    Photo of Francois Garillot

    Francois Garillot


    François Garillot is a data scientist at Swisscom, where he works on curating and understanding telecommunications data through big data tools. Previously, François worked on Apache Spark Streaming’s reliability at Lightbend (formerly Typesafe). His interests include machine learning—especially online models, approximation and hashing techniques, control theory, and unsupervised time series analysis—skiing, sailing, and hunting for good cheese.

    Comments on this page are now closed.


    Picture of Francois Garillot
    09/29/2016 5:21am EDT

    here’s a link to the slides :

    Don’t forget to vote if you liked the talk :)

    09/28/2016 12:56pm EDT

    Can you share the link to your slides.