A datathon will be held on the weekend of 7th – 9th  of November and will be linked to a conference on the Icelandic data ecosystem (GAGNVIST) which will be held on Thursday, November 27th. The winning solution of the datathon will be presented in a session of the conference.

Datathon - Project Description

Figure 1. Illustrates the various factors affecting export dependencies. These factors can be used as key inputs for AI deep learning prediction models.

Datathon Overview

The goal of the datathon is to provide participants with hands-on experience using high-value datasets (HVDs) for practical applications. Participants will develop realistic solutions addressing the economically significant challenges of Icelandic exports of products and services[1], which are influenced by both national and international factors.

Usecase Task Description

Participants (“hackers”) will apply artificial intelligence (AI) techniques—particularly in the area of deep learning (DL)—to train a model using historical monthly data on Iceland’s export values across a wide range of products and services. The objective is to create a forecasting model capable of predicting short- and long-term export trends. To improve model accuracy, participants are encouraged to incorporate additional economic indicators such as inflation[2], cargo flight numbers, and exchange rates, which significantly influence export performance (cf. figure). The purpose of the datathon is to identify the most effective forecasting model and to analyze the effects of key influencing factors.

Data Access

Participants will have access to all relevant datasets through a shared online directory. The dataset will include historical export data as well as economic indicators such as inflation, exchange rates, and cargo flight statistics.

Target Participants

The datathon is open to university students in computer science, engineering, or related disciplines, as well as SMEs and other interested individuals. Participants should have practical knowledge of programming (e.g., Python), deep learning, and data processing/management.

Preparation 

Preparatory presentations, technical instructions, and details on data access will be provided in late October.

For further information contact Seyedreza Hassanianmoaref (seh@hi.is) and Ingólfur Hjörleifsson (ingolfuh@hi.is).

[1] This may refer to total exports from Iceland, or a more granular breakdown — such as exports to one or more specific countries, exports of a particular product by country, or total exports of that product.

[2]  There is a difference between Consumer Price Index and Inflation. See it here