08 Apr


Lifelogging for insights: Applying data science to user behaviour data

AW Seminar



Sony recently launched LifeLog API for application developers. Applications can now access users’ lifestyle, fitness and health data, collected through smartphone sensors and connected SmartWear devices. In this talk we will show how to derive insights from individual user’s LifeLog data using machine learning. Especially, we will look at how to derive personal points of interests from location traces to make your application contextual. Scaling machine learning to millions of users, requires a different approach. We will take a look at how we at the Sony Lifelog team use Spark to scale analytics to millions of users, and how we use it to create new features for Lifelog users, through some live coding.


Håkan Jonsson started out in research in genetic programming for financial forecasting and evolutionary computation, but has since then been developing telco and mobile applications and software for 15 years as programmer, product manager, CTO in startups as well as established companies. For the past 7 years he has returned to research, at Sony Mobile Communications and Lund University, with context awareness, mobile sensing and social computing as his subjects.


Tags architecture lifelogging machine learning mhealth scaling

Platinum Partners


Premium Partners

Verisure100 Netset100px Qlik100px Axis100px Neo4j 100px Cenito100px Mediaevolution100px Ecraft100px Tretton37px100 Combitech100px Playtech100px Webstep100pxnew Securelink100px


Simplyawe100px Crosstech100px Empear100px Oneagency100px Tieto100pxnew Vergic100px Handelsbanken100px Lytics100px Softhouse100px Thirdbase100px Purplescout100px Apsis100px U blox100px Lundalogik100px 24hr100px Schneider100px Yrkesorange100px Experis100px Oddhill100px Se 100px Hotnew100px Tedkomp100px Stratiteq100pxbl Hero gaming logo web partner 100 Seavus web partner 100 Frontiot100px Af100px Ifacts logo partner 100px Resursbank 100px Falconio100px Pinmeto100px Additude100px Bimobject100px Stickybit100px Terranet100px Humanit100px
Sign in