340 likes | 344 Views
Explore the power of big data and the paradigm shift it brings to computational social science. Discover the interconnectedness of people through short chains of acquaintances and the scale at which data can be analyzed. Uncover the potential of computational statistics and the ethical issues that arise. Adapted from H.G. Wells' "The War of the Worlds".
E N D
BIG DATA in a small world Alessandro Panconesi Informatica
Outcome 30% of the letters reached the target Average (median) chain length was roughly six
Conjectures Take any two people in the world, and they will be connected by a very short chain of acquaintances
A matter of scale 300 60,000
A matter of scale 230,000,000 300 60,000
A matter of scale 230,000,000 300 750,000,000 60,000
A matter of scale 230,000,000 300 750,000,000 60,000 Non trivial computational issues
Google Flu Trends Nature 457, 1012-1014 (2009)
More Evidence Movies Hits Stocks? …
No one would have believed in the first years of the XXI century that this world was being watched keenly and closely by awesome computing and financial powers; that as people busied themselves about their various concerns they were scrutinised and studied, perhaps almost as narrowly as a person with a microscope might scrutinise the transient creatures that swarm and multiply in a drop of water. … Adapted from “The War of the Worlds” by H.G.Wells
All'inizio del XXI secolo nessuno avrebbe mai creduto che il nostro pianeta fosse sotto osservazione da parte di entità cibernetico-finanziarie di inimmaginabile potenza e che mentre l'umanità si affaccendava nella sua quotidianità i suoi movimenti venivano osservati da vicino da occhi cupidi, allo steso modo in cui i batteri che pullulano e si moltiplicano sono studiati attentamente al microscopio… Adapted from “The War of the Worlds” by H.G.Wells
FINAL REMARKS • Paradigm shift in the social sciences? • Computational Statistics (more predictive power) • Hybrid science: data processing CANNOT be naïve • New algorithms are the key • Ethical caveats apply (privacy etc)