Every day we generate more than 2 x 10^30 bytes of data coming from a variety of different sources, from sensors to web sites, from transaction records to digital pictures. This data are no longer manageable with known data management tools and data processing techniques. On the other side, enterprise performance management measurement as well as strategic decision making require the capability of implementing the best analysis in a very strict timeline. Thus, an unexpected number of business applications are available for enterprises and in the very next future the competition in several sectors will be determined by new capabilities in properly analyzing data.
The School of Management and Decisions Sciences has planned for this academic year a series of lectures held by distinguished speakers at international level. These lectures, taught by great worldwide recognized experts, will give a unique opportunity to executives and practitioners to be aware of the up-to-date management methodologies and tools.
The Best Ways to Learn from Data
We introduce the “art” of learning from data offering an overview of the modern statistical techniques. The course will offer practical advises as well as methodological insights for social scientists and other applied statisticians. It will provide guidance into the process of building and evaluating models. Data analysis, variables transformations and graphical summaries of the results are emphasized.
We finally discuss some challenges that researchers have when trying to learn from real big data. We then present, through examples, various methods we use in our research.
Andrew Gelman is a professor of statistics and political science and director of the Applied Statistics Center at Columbia University. He has received the Outstanding Statistical Application award from the American Statistical Association, the award for best article published in the American Political Science Review, and the Council of Presidents of Statistical Societies award for outstanding contributions by a person under the age of 40. His books include Bayesian Data Analysis (with John Carlin, Hal Stern, and Don Rubin), Teaching Statistics: A Bag of Tricks (with Deb Nolan), Data Analysis Using Regression and Multilevel/Hierarchical Models (with Jennifer Hill), Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do (with David Park, Boris Shor, Joe Bafumi, and Jeronimo Cortina), and A Quantitative Tour of the Social Sciences (co-edited with Jeronimo Cortina).
Andrew has done research on a wide range of topics, including: why it is rational to vote; why campaign polls are so variable when elections are so predictable; the statistical challenges of estimating small effects; social network structure; arsenic in Bangladesh; radon in your basement; toxicology; medical imaging; and methods in surveys, experimental design, statistical inference, computation, and graphics.
Day one (Oct, 23)
|15:10-17:00||Regression models for prediction|
|17:00-19:00||Statistical modeling and statistical practice|
Day two (Oct, 24)
|15:00-17:00||Statistical modeling and statistical practice|
|17:00-19:00||Two more complicated applications|
Day three (Oct, 25)
|15:00-17:00||Generalizing from sample to population|
|17:00-19:00||Checking and understanding fitted models
Click here to download a copy of the slides used during the course.
Regression models for prediction
- Linear regression
- Logistic regression
- Choice models
Statistical modeling and statistical practice
- Of beauty, sex, and power: statistical challenges in estimating small effects
- Why we (usually) don't care about multiple comparisons
- Some recent progress in simple statistical methods
Information visualization and statistical graphics
Two more complicated applications
- Hierarchical modeling and prior information: An example from toxicology
- Getting around detection limits in diluition assays
Generalizing from sample to population
- Multilevel regression and poststratification
- Survey weighting is a mess
- Causal inference
Checking and understanding fitted models
- Simulation-based model checking
- Tools for understanding complex models
Checking and understanding fitted models
A recording of the module Regression models for prediction is available on YouTube at the following link
|Tou Yun Xiang Pietro||Accenture|
|Abete Tiziana||Banca d'Italia|
|Cascarano Michele||Banca d'Italia|
|Crispino Paolo||Banca d'Italia|
|Di Salvatore Antonietta||Banca d'Italia|
|Matteucci Mirko||Banca d'Italia|
|Paduanelli Stefania||Banca d'Italia|
|Paparo Claudia||Banca d'Italia|
|Scambelluri Valerio||Banca d'Italia|
|Vercelli Francesco||Banca d'Italia|
|Gaudio Rossella||Consorzio per il Distretto dell'Audiovisivo e dell'ICT|
|Milizia Francesco||Regione Lazio|
|Guarini Maria Rosaria||Sapienza - Università di Roma|
|Tognetti Gian Roberto||Sapienza - Università di Roma|
|Brunetti Paolo||Thrust Direct|
|Federico Bruno||Università Di Cassino E Del Lazio Meridionale|
|Abate Giorgio||Università Di Roma Tre|
|Pierini Andrea||Università Di Roma Tre|
The course will be held in "Aula Gini".
Facoltà di Ingegneria dell'Informazione, Informatica e Statistica
Sapienza – Università di Roma Piazzale Aldo Moro 5 - Roma
(Click here for information on how to reach the room)
Course attendes will be provided with wireless Internet access through the Sapienza Wireless service.
€600 for all the course. A 50% discount is available for PhD students, master degree students and post-doc fellows. Please, contact the course secretariat if you wish to apply for a discount.
To subscribe to the course please follow the instructions reported in the registration form. The submission deadline for registration requests is Monday, October 21 16:00. Please, contact us in case of late registration.
For information on programme, costs and discounts contact the course secretariat:
Sig.ra Cristina Puteo