Topics: Coronavirus | AI | Data Science | Deep Learning | Machine Learning | Python | R | Statistics

KDnuggets Home » News » 2013 » Jun » Competitions » VAST Challenge 2013: Predicting movie box office openings and movie ratings ( 13:n15 )

VAST Challenge 2013: Predicting movie box office openings and movie ratings


This challenge focuses on predicting movie box-office openings and movie ratings. Submission deadline is July 8, 2013.



VAST 2013 Challenges, held as part of the IEEE VIS 2013 conference (Atlanta, GA, Oct 13-18, 2013) have 3 different tasks:

  • the boxoffice prediction - described below
  • a large scale network monitoring design problem
  • detecting unusual happenings on large scale network data.

VAST Challenge 2013 Mini-Challenge 1Boxoffice VAST Challenge 2013concerns predictive visual analytics.

This challenge will be a rolling competition with multiple opportunities to participate and ever-changing datasets reflecting the different movies.

The theme is movie success at the box office and in viewer ratings. Participants will be asked to predict how well a set of movies will do at the box office in terms of box office "take" (ticket sales) and how well they will do in the eyes of the viewers (the movies' viewer ratings) for their opening weekend in the U.S. A key feature of the challenge, though, is that contestants will use visual analytics to support their movie analysis and show us how it was used in their analytic processes. So, while the two numeric predictions would be possible to provide by plugging lots of data into a model, we will ask some additional questions to go along with the predictions that will require a human-in-the-loop and hopefully some outstanding visualizations.

Data

Another new feature of this mini-challenge as compared to previous years is the data that contestants may use. The data will still be "closed world", that is, you must use the data we specify and no other. This data will not be provided all at once as in previous years. We will be using data from three sources: IMDb, Twitter, and bitly (bitly data is forthcoming). When you register to participate, you will be required to confirm that you (and any teammates you have) will follow all the terms and conditions of this Challenge, plus also those of IMDb, Twitter, and bitly. More information on the data can be found on the Data webpage for this site. Each of these data sets will be periodically updated. The IMDb data will be updated each Friday. Twitter and bitly data concerning a movie will be updated each day from the day the movie is announced as being part of the Challenge, through the Thursday before the opening weekend.

Registration

All participants for VAST Mini-Challenge 1 this year must register. This includes providing your name and contact information, plus agreeing to the terms and conditions for the use of the data. Please see the registration page for details.

How the Challenge generally works

Movies in the U.S. typically open on a specific weekend. We will select one or more movies for certain opening weekends about two weeks in advance. Data will be made available on those movies for you to use in making predictions. You will use your interactive visual analytic system to make your assessment, and submit both the predictions and a snapshot or two showing how you arrived at them. Submissions are due the Thursday just before the opening weekend at 5PM Pacific time (Friday 1AM GMT). You may send a submission earlier, and at some point we will be giving credit for early submissions. Postings after the submission date and time will not be scored.

Movie goers will then flock to the movies that weekend (or not).

On Monday after the opening weekend, we will post some preliminary figures on box office take and viewer rating. On the Friday following the opening, we will post official numbers as reported by IMDb. We will calculate how close your predictions came and also assess the visualizations you submit. Recognitions will be posted on the leaderboard.

Participate in this challenge at

boxofficevast.org/


Sign Up

By subscribing you accept KDnuggets Privacy Policy