/ CLAUDIO BELLEI
/ CARLO BERGOGLIO
/ MATTEO CHINAZZI
/ LUCA FERRERI
/ MELTEM GÜREL
/ GORDON HATUSUPY
/ BRYAN IOTTI
/ MARCIN KASZYNSKI
/ VITO MANDORINO
/ VAHAN NANUMYAN
/ LORENZO NAPOLITANO
/ ORION PENNER
/ PIERRE PRINETTI
/ CARLO RECCHIA
/ ANNA ROGINSKA
# Data Science
Andrè Panisson (www.isi.it)
Alina Sîrbu (Alina Sîrbu LinkedIn)
Laetitia Gauvin (www.isi.it)
Luca Maria Aiello (www.lajello.com)
Joachim De Beule (engagor.com)
Bruno Gonçalves (www.bgoncalves.com)
BIG DIVE 3 WEEKLY PROGRAM
THIRD EDITION FINAL PROJECTS
The project is aimed at analyzing the iCoolhunt dataset in order to find possible correlations between color space and emerging trends.
Through a filtering system, based on tag and categories, the users can discover the relationships between the images.
The developed platform allows also to upload a new image and to analyse it by comparing its parameters with other images in the same category.
The project analyzes the open dataset (available on dati.camera.it) containing data from Italian chamber of deputies (people, mandates, votes, etc…).
The model shows the vote patterns in order to compare group affiliation and party strategy with the actual behavior of politicians.
The final goal is the search for a possible predictive model based on the vote analysis, that could be used for example to anticipate government crisis.
The group work is focused on two main goals:
– creating an ontology based on color psychology;
– building a platform for the exploration of the iCoolhunt images.
The users may explore the images through clusters and filters in order to have a detailed analysis of tags and most dominant colour for each image. Furthermore the model provides a kind of prediction for the semantic attributes related to the pictures.
GitHub repository / Video
The core concept of the project is the importance of “networks” for a successful startup.
The data analysis allowed to investigate the geographical relationships. The visual representation shows firstly an overview about the Italian startup scene (the cities with the largest number of startups and the relative average revenues) and then it allows to analyze in details how the startups are linked to other cities.
Presentation / Video
The group work focused on using the dataset from GetItalia to track vehicles (mainly trucks) mobility over the Italian territory. The two main goals are:
– identifying more used paths. A practical usage of this information would be a kind of truckpooling or path optimization for the shipment companies;
– identifying lack in gas station coverage over the territory based on the stop (for refueling) analysis.
According to the model developed a kind of prediction system could be implemented.
GitHub repository / Video