For this project commissioned by the GoShort Festival we were asked to create an interactive and intuitive version of the festival programme on large touchscreens.
All the shorts films were given their own identity by generating unique slitscans from the movie's frames. This method gives the viewer an impression of the length, colors and atmosphere of the film.
The programme consists of movies grouped in themed blocks. By layering all the slitscans for every block it’s easy to get an overview of the atmosphere of the complete block without having to leave the main interface layer. If you are interested in a block it’s just a matter of tapping the block and all the movies will pop out.
Simple algorithm for crowd simulation inspired by praying muslims at the Tahrir Square in Egypt. After seeing these beautiful patterns emerge from thousands of people simultaneously performing a single ritual I was interested in figuring out how these patterns developed.
As seen in the image above the praying people align themselves relative to each other forming wavy rows. These rows diverge and different rows begin to form in between those rows. This algorithm is a study into trying to dissect this behavior into simple rules. Once all rules are defined we can start simulating agents that follow those rules.
Rules derived from the picture above:
Agents can only only be placed relative to another agent (cohesion)
Agents prefer to be placed alongside each other to form rows. Only a small percentage of agents gets placed above or below their spawn point (forming rows)
The orientation of agents is slightly offsetted from their neighbors (wavy rows instead of straight lines)
The video above shows a debug view of the simulation. Red areas around agents mean that there's room for another agent in that direction. This information is used for rule 1 and 2.
Through the simulation of hundreds of extremely simple agents seemingly complex behavior can simulated. As seen from the final result below: agents match the behavior of the people seen in the picture.
A concept for a future global energy management scenario where every city generates it's own electricity. If the city has overcapacity it can distribute to the 10 nearest cities to catch spikes. Cities can also request power from neighboring cities if their own demand exceeds their own supply. By linking cities around the globe we end up with a decentralized free-flowing energy distribution system that can absorb outages and other catastrophes.
I developed an application to see the concept in action. The gray band moving from right to left represents the day/night cycle. At daytime a city will consume more power and when night falls the consumption will decrease. You can zoom and pan on the map. Space toggles a world map (internet connection required).
While researching UNESCO's list it appeared that a lot of world heritage sites are related to other geographical features. As an example the "Frontiers of the Roman Empire" was listed as a single coordinate. However it's description is referring to 10 different locations all around Europe. By showing all of these locations a much better impression of the actual scale and size can be given.
After finding a large number of sites that were in need of multiple geographic references this interactive map was born.
This project aims to explain the inner workings of the BitTorrent protocol. When you create a torrent from files those files get chopped up in to a hundreds of chunks. Those chunks are distributed across all peers that are downloading the same file. Because the torrent's contents are scattered across all downloaders your computer has to request and reassemble all these chunks back into the original files.
With this installation I'm visualizing the whole process behind downloading a torrent. All the peers—people downloading the same torrent— are displayed on a world map. All chunks—a small part of the torrent's contents—are displayed on a big computer screen. Any connection that transmits data between the peer and the computer is represented by a line from the peer's location to the actual chunk that's being downloaded.
The video below is an impression of the installation in action. This work was on display at the ArtEZ 2012 graduation show.
A small project I did for our graduation show. I've used 3d models scanned from my fellow student's faces to create a 3d point cloud viewer. We used this program to display our portraits in a fresh, unconventional new way.
The final result is a projection that was installed at wall of our graduation show.
In 2006 AOL accidentally released a text file containing twenty million search keywords for over 650,000 users over a 3-month period. The data was anonymized however the queries were accompanied by a unique numerical ID.
All the search queries in this project originate from one single user. I've collected all his search queries and categorized his queries in 8 categories. Because this user is very verbose we get a pretty good insight in what is going on in his life. He's dealing with a sagging roof, he has some problems with an inheritance, copd, his sick dog and a bunch of other things.
I've created an interactive calendar where you can read all of his questions on a chronological timeline. Enable or disable highlighting of categories to follow different story lines.
A series of data visualizations for a generative publication that can be published weekly without human intervention: the Wikipedia vandalism report.
Through the Wikipedia API all modifications to all wikipedia articles were stored for automatic processing. An algorithm was created to determine if an edit was malicious or genuine.
Every week the algorithm ran over the data collected in the previous seven days to determine which articles suffered most from online "vandalism". Reflecting world-wide sentiments on subjects like the news and pop culture. Articles on the "Tea Party movement", "Dancing with the stars", "The Notorious B.I.G", "Fred Phelps" and Wikipedia's own "Administrator intervention agains vandalism" were top ranking vandalism targets.
The graph above shows 7 days of heavy modification on the Tea Party movement article. Users are listed on the right, individual edits are represented on the left. The length of each bar represents the total length of the article after the edit has been applied.
Distribution of malicious (red) and regular edits (green) on an article about Ei-ichi Negishi, a Japanese chemist. This graph was generated right after he won the 2010 Nobel Prize in Chemistry. Apparently reason for both positive and negative attention.