AI2S Winter Datathon 2021: A 24-Hour Data Science Open Challenge!

 

TOPIC OVERVIEW: CRYPTO ART

Crypto Art is a rising art movement that associates digital artworks with non-fungible asset tokens. It draws its origins from conceptual art, sharing the immaterial and distributive nature of artworks, the tight blending of artworks with currency and the rejection of conventional art market and institutions. Similar to traditional artwork, the concept allows you to buy, sell and trade digital goods as if they were physical. Collectors of Crypto Art are overwhelmed with choices, reason why matching consumers with the most appropriate artworks is the key to enhancing user satisfaction.

PROBLEM STATEMENT: CAN YOU SUGGEST AN ARTWORK TO AN ART COLLECTOR?  

Given a set of features of artworks and a set of buy/sell transactions made by several collectors, the goal is predicting the interest of the different collectors to the related art pieces. Different methods can be used to solve this problem (e.g. recommender systems) but any creative solution is welcomed and encouraged!

EVENT TIMELINE: SUBSCRIPTION & SUBMISSION DEADLINES

The deadline for the registration is at 11:59 PM on Feb the 25th. The contest will begin at 10:00 AM on Feb the 27th and it will end at 10:00 AM on Feb the 28th. The deadline for submitting the solution coincides with the conclusion of the contest. The time offset considered is UTC+01:00.

MAIN SPONSOR: ESTECO

ESTECO is an independent software provider, highly specialized in numerical optimization and simulation data management with a sound scientific foundation and a flexible approach to customer needs. ESTECO's technology brings modularity, ease of use, standardization, and innovation to the engineering design process. ESTECO's smart engineering suite brings enterprise-wide solutions for design optimization, simulation and process data management (SPDM), and process integration and automation. With 20 years' experience, the company supports over 300 leading organizations in designing the products of the future, today.

DATASET

You can find the dataset at https://github.com/demirbilek95/AI2SWINTERDATATHON_PROBLEM

NEWS & UPDATES

The teams who officially subscribed to the event can be found here. 
Introduction to the event and explanation of the problem can be found here.
The slides of the presentation can be found here.

View full rules

Hackathon Sponsors

Prizes

€500 in prizes

1st Place

The best overall project taking into account all the judging criteria.

Devpost Achievements

Submitting to this hackathon could earn you:

Eligibility

Participants that do not comply with the following requirements may be rejected or disqualified by AI2S in its sole discretion:

  • Participant must be at least 18 years old, and in their majority according to the jurisdiction of the country of residence of the participants at the start of the datathon.
  • Organizers, volunteers, judges, sponsors, or in any other privileged position at the event are not eligible to participate in the datathon.

Requirements

WHAT DO I NEED TO DO TO PARTICIPATE?

AI2S Winter Datathon is a team-based competition. Entrants can build a team, composed of one to five members, or participate singularily (they will be considered as a single member team). The enrolment to the contest consists of two compulsory stages, to be strictly done in the following order:

  • Entrants eligible for the competition must create an account on Devpost and register to the event.
  • Teams must choose a Team Leader (which coincides with the single participant in case the team is composed by a single element) who is the only member of the team who must fill in the google form, by the end of the subscription deadline (25th of February).

Once correctly concluded the latter steps, entrants will be officially considered as participants to the event. 

WHAT DO I NEED TO BRING FOR SUBMISSION?

To qualify for prizes, participants must respect the following submission requirements:

  • Team Leaders ONLY have to submit their solution to Devpost platform (according to the Devpost's requirements) by adding a ZIP file taking the name of their team and which must contain the following elements:

    • A CSV file containing the prediction of the target variable (named as prediction.csv).
    • A PDF file, composed of maximum three pages, containing the logical process adopted for the resolution of the problem, having as font the Times New Roman of size 12pt (named as report.pdf).
    • A FOLDER containing the developed code capable of reproducing the prediction (named as src). The FOLDER should also contain a description of the resources needed to correctly create and/or execute the executable code. The internal structure of the FOLDER is at the discretion of the participant.
       
      A visual representation of the above can be found at the following github repository.
  • Participants must develop the code using only open source software, licensed under an Open Source Initiative approved license (opensource.org/licenses), and are not allowed to use any other data sources than the dataset released by AI2S at the beginning of the datathon.

Based on the prediction.csv files received from the teams participating at the competition, a Top 10 Ranking will be created using the RMSE (Root Mean Square Error) score (the less the better). Only submissions within the Top 10 Ranking will be evaluated by the judges and will compete for the final prize. 

 

Judges

Eric Medvet

Eric Medvet
Associate Professor at University of Trieste

Marco De Pasquale

Marco De Pasquale
Software Developer at Esteco

Marco Pividori

Marco Pividori
Software Developer at Esteco

Luca Manzoni

Luca Manzoni
Assistant Professor at University of Trieste

Judging Criteria

  • Clarity
    Has the candidate expressed his/her proposed solution in a scientific and procedural manner (report.pdf)?
  • Motivation of Choice
    Has the candidate justified his/her choices with concrete hypothesis (report.pdf)?
  • Code Quality
    Has the candidate made it easy to read and understand sections of the code (src)?
  • Accuracy
    The accuracy of the predictions will be evaluated using the root-mean-square error, or the square root of the quadratic mean of the differences between the predicted and observed values (prediction.csv).