PRC Data Challenge 2026

PRC Data Challenge 2026 banner

The Performance Review Commission (PRC), with the support of OpenSky Network (OSN), invites data scientists, aviation professionals and data analysts to collaborate and compete in a data challenge to develop innovative methodologies for benefit of the broader aviation community.

The aim is to do this through:

  1. Using open data,

  2. openly sharing final solutions (incl. code and documentation on a public GitHub repo),

  3. publishing an open access paper about the implemented solution on Journal of Open Aviation Science (JOAS) .

Data Challenge

The 2026 edition of the PRC Data Challenge invites participants to predict the taxi-out time of flights from 10 major European airports.

Scope

The learning dataset will consist of all 10 airport movements for the full year of 2025.

Ranking

The applied ranking will use Root Mean Square Error (RMSE) for submitted predictions for the movements in January and July 2026.

Timeline

The competition run from the \(1^{st}\) of September 2026 up until the \(14^{th}\) of October 2026 12:00:00 CEST (2026-10-14T10:00Z).

The final phase with 2 additional months to predict, Feb and Jun 2026, was introduced on 8th October. The goal was to avoid overfitting or reverse-engineer the ranking algorithm. The final set of predictions was not part of the ranking leaderboard.

What’s next

Data preparation:

  1. collect all <team-name>_final.parquet submissions uploaded before the deadline (or equivalently saved submissions in the team’s GitHub repo)
  2. fork all shared repos under a dedicated GitHub organisation and tag the last commit before the deadline

Evaluation:

  1. rank the submissions
  2. evaluate code and documentation at the tagged commit

Final ranking:

  • combine rank and documentation in a final leaderboard.
  • present the Challenge and announce the winners at the 14th OpenSky Symposium held at CRIDA in Madrid, Spain on 29-30 October.

Dissemination:

  • share sympoium slides and material on 2026 challenge website (Dec 2026)
  • pubblication of data paper on JOAS describing the dataset of the challenge (< end of Q1 2027), and
  • pubblication of dataset of the challenge as open data (simultaneously with data paper)
  • acceptance and review of JOAS papers describing their model by participating teams (Q1 2027)
Reward

The combined prize amount for the first 3 teams is 5000 EUR.

Terms & Conditions

Check the Terms & Conditions for eligibility for participation and prize before filling in the team creation request form.

NotePlease note

We reserve the right to change the rules or even stop the challenge due to any circumstances.

Participate

Join the competition; request the creation of your team here!

Request Team Creation!

ImportantImportant

Your request will be assessed by the PRC Data Challenge team for approval. Upon approval you will receive a verification email and, on successful reply, the relevant access keys to the datasets/buckets.

Note: The participation is open to all, but teams from sanctioned countries are excluded.

Social Media Interactions & Forum

Join the conversation and stay up to date with the Data Challenge news and announcements on the OpenSky Network Discord channel #prc-data-competition .
Note: you need to have access to the OpenSky Network Discord server first.

Outreach

We are strong advocates of open science and (strongly 😉) encourage teams, particularly the top-ranked teams, to submit a paper to JOAS describing their approach and the inner workings of their models.
Here are some examples from previous editions:

We also publish a data paper describing the challenge dataset, with the aim of making the data more accessible and supporting further use in research, education, and other applications.
Examples from previous editions include:

The following GitHub organizations host the repositories of participating teams:

List of Acronyms

JOAS: Journal of Open Aviation Science

OSN: OpenSky Network

PRC: Performance Review Commission

RMSE: Root Mean Square Error