Hintergrundbild © Stuttgart-Marketing GmbH

BTW 2017 Data Science Challenge SDSC17

The Data Science Challenge has finished. The theme was: How can we avoid car accidents?

1st Place: Ghareeb Falazi, Majd Abdo, Yuliya Volga

2nd Place: Alexander Askinadze, Matthias Liebeck

3rd Place: Tobias Hildebrandt, Conrad Indiono, Prof. Stefanie Rinderle-Ma

3rd Place: Martin Grimmer, Matthias Kricke, Eric Peukert



Primary data set

The following data sets are optional

Exploration

  • Visualize the data and reveal underlying patterns like:
    • Are there accident-free/exceptionally safe roads?
    • Where are the most dangerous spots?
    • Show the development over time in an animation.
  • Can you identify classes/types of accidents? Which ones are the most frequent?

Analysis

  • What factors have the biggest influence on accident probability?
  • Can you observe an unusual/significant increase or decrease in accidents anywhere? What actions or events caused it?
  • Can you find a connection between accidents and large public events?
  • Propose actions that will lower the number of accidents based on your results from above:
    • What could the city change or improve?
    • What should drivers do differently?
    • Are there places where drivers get blinded by light regularly?
    • ...
The jury will decide on the winner based on the following criteria:
  • Impact and novelty of the results. How large is the potential for avoiding accidents?
  • Comprehensiveness and scope of the results.
  • Visual/aural representation of the results.
  • Live-presentation at BTW on March 7th.
You can do either a comprehensive but shallow analysis, or a narrow and deep analysis that focuses on one question.

For any questions please contact: sdsc17@ipvs.uni-stuttgart.de



Demonstrate what an insightful data analysis or visualization you can create using your own approach, and win a prize!

Take the opportunity to create your own cloud-based data analysis approach and compete with other contestants. You can present your results during the BTW2017 conference in Stuttgart in front of an expert panel from industry and academia who will choose the best contributions. The winners will be awarded a certificate as well as prize money.

Prizes

  • First place: 500 Euro
  • Second place: 300 Euro
  • Third place: 200 Euro

Challenge Structure

We publish preliminary, exemplary data sources as well as analysis tasks with this challenge announcement. You can use this preliminary data and tasks in order to plan and prepare your approach. You can then apply for the challenge by submitting a two-paged proposal. We will notify the accepted proposals and publish the final challenge data sets and task one month before the BTW2017 conference. These final data sets and task will be similar enough to the preliminary ones so that you can adapt your approach to them. The actual challenge is therefore to adapt, modify, and finalize your approach within one month in order to solve the final challenge task. So it is useful to choose an approach that is flexible and reusable.

The goal of the challenge is to integrate and analyze data sources. The result can be an actionable insight or insightful visualization. Contestants can freely choose the cloud platform and technology they prefer. Your approach may integrate existing tools and services, or create something new. The data for the challenge will be accessible via the cloud platform IBM Bluemix. As a sponsorship, you can use Bluemix, and its analytics services, for free in the context of this challenge. If you're interested in free access to Bluemix, please contact sdsc17@ipvs.uni-stuttgart.de or state so when submitting your proposal.

The Data Science Challenge is open for PhD, masters, as well as bachelor students. Individuals as well as teams may participate. At least one person per team has to register for the BTW in order to present the results during the conference. This registration is free of charge for students participating in the challenge.

Schedule

  • Publishing of the preliminary data sets and tasks.
  • Applications are open. Submit a two-paged proposal explaining your approach and used technologies. State your relevant prior experience in the field.
  • 31.10.2016: Applications are closed. (Deadline extended)
  • 30.11.2016: Notification of accepted proposals.
  • 06.02.2017: Publishing of the final data sets and challenge task. From now on you have one month to adapt, modify, and finalize your approach.
  • 07.03.2017: Presentation of the results at BTW2017 in Stuttgart.
  • 09.03.2017: Ranking of the contributions and awarding of the winners.
  • 13.03.2017: All participants receive an invitation to submit a scientific article discussing their approach.

Preliminary Data Sets and Tasks

Use the following preliminary data sets and tasks in order to prepare your approach and proposal. The final challenge data sets and task will be published one month prior to the BTW2017 conference. The actual challenge is therefore to adapt, modify, and finalize your approach within one month in order to solve the final challenge task. We chose these preliminary data sets and tasks in such a way that you can reuse and adapt your approach to the final challenge data sets and task.

Main Data Sets:

Additional Data Sets:

Tasks:

  • Find interesting patterns and behaviors. Visualize your results.
  • Predict when and where the demand will be the highest.
  • Predict the "User Type" based on the given data.
  • Generate recommendations for improving the infrastructure.

Proposal Submission

Submit a two-paged proposal (in german or english) explaining your approach and used technologies. State your relevant prior experience in the field.

Use the LNI template for your proposal: See submission page.

Submit your proposals via CMT: https://cmt3.research.microsoft.com/BTW2017 (Data Science Challenge Track)

For any questions please contact: sdsc17@ipvs.uni-stuttgart.de

Good Luck!

Sponsors

Platinum

Silver

Supporters

© 2016 - 2017 Universität Stuttgart