Social Mining & Big Data Ecosystem H2020 - www.sobigdata.eu September 2015- August 2019 SoBigData meets EUI -– 11 th October 2017 @SoBigData (h7ps://twi7er.com/SoBigData) h7ps://www.facebook.com/SoBigData SoBigData@EUI – 11th October 2017
SOBIGDATA VISION AND VALUES SoBigData@EUI – 11th October 2017
Big data “proxies” of social life Shopping paJerns & lifestyle RelaKonships & social Kes Movements Desires, opinions, senKments SoBigData@EUI – 11th October 2017
Social mining: making sense of big data to understand society SoBigData@EUI – 11th October 2017
What is Social Mining • Automated discovering pa7erns and models of human behaviour across the various social dimensions that have big data “proxies” – desires and opinions – relaKonships and social Kes – life-styles – mobility SoBigData@EUI – 11th October 2017
How opinion emerge and polarize with social media data SoBigData@EUI – 11th October 2017
h7ps://www.buzzfeed.com/jamesball/3-million-brexit-tweets-reveal-leave-voters-talked-about-imm?utm_term=.jmDQE9JNR#.fuOOrb145 SoBigData@EUI – 11th October 2017
SoBigData@EUI – 11th October 2017
EsWmaWng traffic fluxes on road network from mobile phone data B H W A C SoBigData@EUI – 11th October 2017
Measuring happiness with twiJer data SoBigData@EUI – 11th October 2017
Measuring wellbeing with GSM data SoBigData@EUI – 11th October 2017
What is percepKon of refugee crises • Internal and external percepKon by country – Index ρ - the raKo between pro refugees users and against refugees users – Red means a higher predominance of posiKve senKment, higher ρ – Yellow means a higher predominance of negaKve senKment, lower ρ - - - + + + (b) Internal (c) External (a) Overall. percepKon. percepKon. SoBigData@EUI – 11th October 2017
What’s special in social mining? • Any data science experiment is composed by: – data acquiring (open data, crowdsourcing, crowdsensing,) – model building (data mining, machine learning, network science, …and very complex validaKon phase), – creaKon of an exploraKon scenario (what-if analysis) (different validaKon seing), • ….similar to many other data-driven science process,… but data are produced by humans SoBigData@EUI – 11th October 2017
What is needed • responsibility and trust – FACT: Fairness, Accuracy, ConfidenKality and Transparency • harness social mining for scienKfic advancement and for the social good – FAIR: Findable, Accessible, Interoperable, Reproducible • responsible open data science SoBigData@EUI – 11th October 2017
Social mining as OPEN SCIENCE SoBigData@EUI – 11th October 2017
SOBIGDATA GOALS SoBigData@EUI – 11th October 2017
SoBigData GOAL is… TO CONSTRUCT THE MulWdisciplinary European Infrastructure on Big Data and Social Data Mining (the Social Mining CERN) providing an integrated ecosystem for ethic-sensiWve scienWfic discoveries and advanced applicaKons of social data mining on the various dimensions of social life, as recorded by “big data”. SoBigData@EUI – 11th October 2017
The pillars for reaching the goal • an ever-growing, distributed data ecosystem for procurement, access and curaWon of big social data, within an ethic-sensiWve context , based on – innovaKve strategies for acquiring social big data for research purposes, – using both opportunisKc means offered by social sensing technologies and – parKcipatory means based on user involvement as prosumers of social data and knowledge. SoBigData@EUI – 11th October 2017
The pillars for reaching the goal • an ever-growing, distributed plaform of interoperable, social data mining methods and associated skills : – tools, methodologies and services for mining, analysing, and visualising complex and massive datasets, – harnessing the techno-legal barriers to the ethically safe deployment of big data for social mining. SoBigData@EUI – 11th October 2017
The pillars for reaching the goal • Building the Social Mining community of scienWfic, industrial, and other stakeholders (e.g. policy makers), SoBigData@EUI – 11th October 2017
The path to achieve the goals • Integrate European (naWonal) infrastructures and centres of excellence in big data analyKcs, social mining and data science 1. Text and Social Media Mining (TSMM) 2. Social Network Analysis (SNA) 3. Human Mobility Analy?cs (HMA) 4. Web Analy?cs (WA) 5. Visual Analy?cs (VA) 6. Social Data (SD) SoBigData@EUI – 11th October 2017
IntegraWng naWonal research Infrastructures SoBigData@EUI – 11th October 2017
The ConsorWum SoBigData@EUI – 11th October 2017
The path to achieve the goals • Grant access (both virtual and trans-naWonal on-site) to the SoBigData RI to mulK- disciplinary scienKsts, innovators, public bodies, ciKzen organizaKons, SMEs, as well as data science students at any level of educaKon. • joint research, and extensive networking and innovaWon acWons SoBigData@EUI – 11th October 2017
SoBigData@EUI – 11th October 2017
SoBigData e-infrastructure • It is powered by the D4Science plaform o Used as a producKon system by other communiKes o Users: +3600 (SoBigData is 16.5%, fast growing), CompuKng: 1500 cores, Storage: 400 TB) • Supports basic services o AuthenWcaWon, AuthorizaWon, AccounWng framework o Resource Catalogue o Virtual Research Environments management framework (VREs) VRE VRE VRE e-Infrastructure Resources SoBigData@EUI – 11th October 2017
SoBigData VREs • City of CiKzens • Tag-me • Societal Debates • Smaph • Wellbeing • MigraKon SoBigData@EUI – 11th October 2017
Exploratories SoBigData@EUI – 11th October 2017
Exploratories Virtual Research Environments tailored on specific where Social Mining is applied • Promotes results sharing among scienKsts and communiKes • Promotes the use of RI through Virtual and TransnaKonal Access SoBigData@EUI – 11th October 2017
Big Data for Societal Debates PolarizaKon, controversy and topic trends on societal debates through social media SoBigData@EUI – 11th October 2017 Lead by Aris Gionis and Dominic Rout
Polarized PoliWcal Debates Monitoring Topics across Time and space SoBigData@EUI – 11th October 2017
Exploratory: Big Data for City of Citizens SoBigData@EUI – 11th October 2017
Social Sensing Real-Kme crisis mapping AvvenuK, M., Cresci, S., Del Vigna, F., & Tesconi, M. (2016). Impromtu crisis mapping for prioriKzed emergency response. IEEE Computer , (to appear in May 2016). SoBigData@EUI – 11th October 2017
EsWmaWng traffic fluxes on road network B H W A C SoBigData@EUI – 11th October 2017
Big Data for Well Being and Economic Performance DeprivaKon Index (in France) predicted with Mobile Phone traces Lead by Peep Kungas SoBigData@EUI – 11th October 2017
Measuring happiness with twiJer data SoBigData@EUI – 11th October 2017
Network effects are relevant in propagation of financial distress SoBigData@EUI – 11th October 2017
Big Data for Migration Studies DeprivaKon Index (in France) predicted with Mobile Phone traces Lead by Peep Kungas SoBigData@EUI – 11th October 2017
Ethics and trust SoBigData@EUI – 11th October 2017
CreaWng a new generaWon of data scienWsts 8 events and 337 trainees SoBigData@EUI – 11th October 2017
SoBigData@EUI – 11th October 2017
Become a SoBigData Supporter Share and make findable your • data science results Become part of the scienWfic • network hJp://www.sobigdata.eu/join-us • SoBigData@EUI – 11th October 2017
Recommend
More recommend
