Rules and Registration

TARGET

SUS5 is open to national and international University students, with competences in Statistics, Data Science and Machine Learning. The participants work in teams made by a minimum of 2 until a maximum of 5 people. The challenge is in English and teams made by a single participant are not allowed.

A maximum of 200 participants are allowed.

The registrations will close on June 9th, 2019 or once the maximum number of 200 participants is reached.

REGISTRATION

Each participant has to register for the event SUS5 using the form provided in this page. To ensure the participation, the form must be completed in all its parts, declaring also the name of the team, the team leader and one team member who will be contacted to obtain, upon acceptance of the applicable terms and conditions of usage, a free license for the use of the platform SAS Visual Data Mining and Machine Learning before and during the data challenge. The teams compete on the same business problem and anonymized dataset provided by Generali.

By registering to the event SUS5 

• Each participant acknowledges that the dataset, including all its copies, reproductions or parts of it, is and will remain exclusive property of Generali.

• Each participant undertakes (i) to keep the dataset strictly confidential and not to disclose it to third parties for the entire duration of the hackathon and for the following five years; (ii) not to make improper use of it; (iii) not to copy it, reproduce it or keep it, in any form and by any means; (iv) to remove it from his/her own machine and from any support and/or material in his/her possession at the end of the competition; (v) not to allow third parties to access, copy or reproduce it; as well as (vi) to grant to Generali the right to know and exclusively implement the proposed solutions for the problem presented.

• Each participant acknowledges and accepts that all results and/or analyzes produced during the hackathon must be considered as the exclusive property of Generali, which will acquire the full, exclusive and undisputed property (with the express exclusion of the moral copyright) together with the related rights provided by the rules, of any rank, international, European, national and foreign in the field of intellectual property and industrial property. To this end, the participant undertakes to transfer definitively to Generali any data, information, source code, in any form, concerning the solutions generated during the hackathon.

Each participant will hold Generali harmless and undamaged against any breach of the foregoing obligations, or otherwise refund Generali from any amount of money paid for any reason due to such violation.

THE CHALLENGE

The main goal of the teams is to develop analytic solutions which are able to accurately predict a specific response variable of interest as a function of other predictors. To do this, the teams are assigned the same business problem and dataset. For a subset of the data (training set), the teams know both the response and the predictors, whereas for the remaining part (test set) only the predictors are available.

During the challenge the teams should develop an appropriate analytic strategy using the information from the training set and apply it to make accurate predictions for the units in the test set.

The whole competition is managed by the Bocconi Data Science Challenges platform where the teams can

1. Download the dataset
2. Submit their predictions for the test set
3. View their partial ranking (computed on a subset of the test set)
4. Upload their final reports
5. Check their final position, once the challenge is closed.

Such a final ranking is based on the last prediction submitted.

MATERIALS AND SOFTWARES

All teams are assigned suitable locations, equipped with charging stations, wifi connection, glass boards and markers. Each team is responsible to bring the hardwares (laptop, tablets, ...) necessary to analyze the data and may use any programming language or software which is deemed appropriate.

This year, one member of each team is eligible to obtain, upon acceptance of the applicable terms and conditions of usage, a free license for the use of the platform SAS Visual Data Mining and Machine Learning before and during the data challenge. To familiarize with the platform, video tutorials can be found at the SAS website. Moreover, a series of webinars, addressed to all the participants, is available here.

Other materials such as books, notes, slides and internet sources are allowed as long as they do not interfere with the proper functioning of the competition, causing an unreasonable overload of technological infrastructure and servers or intercepting operating systems, data or personal information or violating, stealing or canceling information or computer correspondence to third parties.

The entire teams activity must be carried out only within the competition venue. The help of external collaborators cannot be requested.

FINAL EVALUATION

For the evaluation, the teams are asked to upload in the Bocconi Data Science Challenges platform their final predictions for the test set and a report describing their final solution along with a discussion on relevant patterns that have been found in the data and that might be useful for the company to address the business problem provided.

These reports:

• Must not be offensive, insulting, defamatory, libelous, pornographic, vulgar, obscene, pedopornographic and, in any case, not in disagreement with the principles of public order or morality or that may cause harm in any way to minors. 
• Must not contain viruses or other programs aimed at damaging or interfering with the proper functioning of the competition. 
• Must not contain advertising, promotional material or any other form of unrequested or unwanted solicitation. 
• Must not concern minors in any way. 
• Are not hand back to the teams at the end of the competition.

Based on the materials provided at the end of the challenge, three winning teams are selected by a committee of Data Scientists from academia and industry. The evaluation and the awards assignment are based solely on merit criteria. These include, but are not limited to

• Predictive performance
• Quality of the final report
• Applicability and effective presentation of the results
• Creativity in combining different analytical methods successfully
• Ability to include external information of public domain
• Capability to use different softwares efficiently

Incomplete or non-compliant materials are not evaluated. The materials received according to times or methods different from those indicated are not evaluated.

PRIZES

Based on the judgment of the committee, which is final, the three winning teams are entitled to receive three different prizes provided by Generali and SAS.

• Visit to Generali Innovation Park and attendance of the Tam Tam Talks along with a Smartwatch for each team member

Prize provided by Generali to the best overall project which combines an outstanding predictive performance and the highest evaluation from the experts’ committee.

• Visit to Generali Innovation Park and attendance of the Tam Tam Talks along with a Bluetooth Speaker for each team member

Prize provided by Generali to the team which provides the solution with the best predictive power.

• Free SAS course for each team member

Prize provided by SAS Institute S.r.l. to the best project performing the analyses with SAS software and platform.

The winning teams are notified in the afternoon of June 19 and are formally announced during the evening of June 19 at the awards ceremony of the SIS2019 conference "Smart Statistics for Smart Applications" (Aula Magna, Università Cattolica del Sacro Cuore, Largo Gemelli 1, Milano). The presence of at least one member of each winning team is required during the awards ceremony (only for the three winning teams).

At the notification step and during the awards ceremony, the winners will receive detailed information on how to claim the prizes offered by Generali and SAS. 

• The organizers assume no responsibility for technical malfunctions that may adversely affect the registration procedure or the notification of the results. By way of non-exhaustive example, the organizers are not responsible for interruptions in network connections; data upload failed, incorrect, incomplete; non-functioning communications with participants due to spam filters that can block communications to participants.

• The organizers are not responsible for the lawful or illicit use of third-party contents used by participants without permission.