Open Data & Tasks

With a view to promote Kobe city where is the venue of SIGGRAPH ASIA 2015, we picked up Scientific Open Data from RIKEN and Kobe University, and Urban Open Oata of Kobe City.

The open data which authors will use is not limited, though we welcome to use following open data.

SSBD: Visualizing Biological Dynamics and Cell Divisions of Worm, Fruit Fly and Zebrafish.

Systems Science of Biological Dynamics (SSBD) database provides a rich set of resources for analyzing quantitative biological data, such as single-molecule, cell, and gene expression nuclei. Quantitative biological data are collected from a variety of species and methods.  These include data obtained from both biological experiment and computational simulation.

There are currently more than 180 sets of 4D quantitative cell division dynamics experimental data on C. elegans (worm) embryo together with some 2.2 million microscopic images stored in SSBD. SSBD also contains experimental data on zebrafish and D. melanogaster (fruit fly) embryos and simulation results on E. coli (bacteria).


New advances in live cell imaging and computer simulation have allowed us to capture and model the dynamical nature of biological development of various species. With the increase amount of images and data, direct manual observation on tracking dynamical behavior has reached its limit. Quantitative analysis is now essential in further our understanding of biology. However, visualization of quantitative data of biological dynamics such as cell division is still in its infancy. Your challenge is to utilize the data in SSBD and to come up with a novel way to display and visualize biological data that will help biologists to advance their understanding.


Task 1: Visualization of quantitative data of biological dynamics and cell division of one or more organisms using BDML formatted data files

Problem overview: BDML can be quite large in size due to large number of cells or large number of time points. Data contain spatial and temporal positions.

Level of difficulty: Medium

Task 2: Visualization of quantitative data of biological dynamics and cell division of one or more organisms using SSBD RESTful API

Problem overview: API provides a way to extract only necessary part of the data. However, searching for data in real time can be quite slow.

Level of difficulty: Medium

Task 3: Visualization of quantitative data of biological dynamics and microscopy images.

Problem overview: Visualizing both quantitative data and image data together can be data intensive.

Level of difficulty: Medium to hard

Task 4: Prediction of RNAi knocked down genes affecting worm development.

Problem overview: Predicting RNAi knocked down genes affecting worm development. Visualization of the predictive results and methods.

Requires some level of understanding of developmental biology.

Level of difficulty: Very hard


  • You can find 136 set of quantitative data of embryogenesis in C. elegans with 72 known RNAi knocked down genes and 50 set of wild type data in SSBD.
    • size of pixel: 0.105 micrometer
    • distance between adjacent focal planes: 0.5 micrometer
    • time interval: 40 seconds
  • You are now given a new set of quantitative data of embryogenesis in C. elegans with an unknown RNAi knocked down gene in the BDML file below.
    • BDML file: RNAiEmbryo.bdml0.18.xml.gz (6.5MB; original xml file 87MB)
    • The new BDML file was taken using slightly different protocols.
      • size of pixel: 0.1015 micrometer
      • distance between adjacent focal planes: 0.5016 micrometer
      • time interval: 20 seconds
  • You are tasked to identify the unknown knocked down gene and to visualize the analytical reasoning of the result.
  • Reference:  Kyoda, K. et al. (2013) WDDD: Worm Developmental Dynamics Database. Nucleic Acids Res. 41, D732–7 [Open Access].


  • Originality: Innovative and creative approach to visualize biological data
  • User friendly/Style: Easy of use and effectiveness in visualizing data.
  • Understanding/Insight: Usefulness in gaining a new understanding of the data.

Contestant can use any open source/commercial tools and programming languages. Extra points will be given to those that can visualize the data on a web browser. Further points will be given to those that can utilize other biological databases.


SSBD web site


  • BDML4DViewer - a plugin of ImageJ software for visualizing BDML
  •  SSBD-OMERO.Insight - a plugin of ImageJ software


For more information, please visit our web site at 


Nowadays, automobile and aerospace industries rely heavily on computational fluid dynamics (CFD) to design and enhance their products. Using CFD techniques, the behaviour of a fluid like air is simulated around and through a given 3-dimensional model of an object like a car, or a jet engine. Normally, the volume around the object is quantized as an array of voxels (3-dimensional pixels) and values like temperature, pressure, and velocity are calculated for each voxel. These values, then, should be translated into visual representation of physical phenomena for the benefit of the engineers and designers so that they can easily see the vital information they need. Participants of SIGGRAPH Asia 2015 Visualization Contest in the category of Computation Fluid Dynamics are challenged to invent, implement, and demonstrate methods for producing visual representation of implicit and explicit information in CDF data with enhanced informative quality.


Each participant team or individual is to choose one from the following tasks:

Task 1. Revelation of Physical Phenomena

A dataset consisting of precalculated results of a CFD simulation is made available to the contestants to download. The simulated volume is 192×64×64 voxels, and the simulation is performed for 50 timesteps. The contestants are asked to devise a software solution to extract implicit information like flow streams or statistical facts from this data and make them visually accessible. 

We are basically asking the participants to extend image processing and pattern recognition algorithms to 3 or 4 dimensions and apply them to temporal and/or volumetric data. Open a pattern recognition or image processing text book, and try to see which set of algorithms can be used to reveal patterns that are hidden or hard to see in the given dataset. Make Bayesian, linear or non-linear classifiers, use clustering, K-means, generate and make use of histogram, create a rule-base, apply fuzzy logic, perform feature selection, apply spacial or frequency domain filters, try morphological filters, and so on, to get a visual interpretation of the data that is both useful and novel.

Task 2. Development of Shader Script
In addition to voxel dataset mentioned in Subject 1, corresponding iso-surface data is also made available for download. Those who pick this subject are asked to use GL Shader Language (GLSL) to enhance the meaningfulness and visual appearance of the iso-surfaces. This may include adaptive application of colors and transparency with regards to the corresponding voxel data, design of an environment shader, and so on. 

To get a better idea about the topic of this category, participants are advised to watch the accompanying tutorials.


The submissions are judged based on 
  • Usefulness, 
  • Novelty, 
  • Clarity, and 
  • Beauty of the output produced by the program. 


  • ParaView Tutorial -- Learn the basics of using ParaView and how to open and visualize our dataset.
  • VTK Tutorial  -- Learn the basics programming using VTK in C++, and how to open and visualize our dataset with it.


  • Subject: Toyota IQ (
  • Credit: CFD simulation data Provided by Dr. Junya Ōnishi from Institute of Industrial Science of Tokyo University using the K-Computer
  • License: Non-commercial Use Only
  • Format: VTK 
  • Download Link: (no longer available 2016/6/2)
  • File Size: 0.98GB 
  • Contents: (1) 3D model of the car, (2) CFD data, (3) Sample videos. 

Watch the following videos to get an idea about the data.

Urban Data

Recently, governments and local governments all over the world publish open government data to enhance citizen's participation and intend to be more open about politics and administrations.
These open government data include much about global economy information, geographic information, and so on.
Visualizations of such open government data (not only them but also any urban data) will help citizens to understand their local or global issues.


We introduce some urban data repositories as follows.
We are welcoming any idea which visualize any urban data.

To help your proposing of visualization, we introduce two TED talks and how to realize the latter example without any programming.

Tim Berners Lee talked about a merit of open data at TED. Let's see this presentation as follows.

Hans Rosling showed a nice example of utilizing open data.

Here we show a same demo which is realized by Google Public Data Explorer.
Google Public Data Explorer makes large, public-interest datasets easy to explore, visualize and communicate. As the charts and maps animate over time, the changes in the world become easier to understand. You don't have to be a data expert to navigate between different views, make your own comparisons, and share your findings. (About the Public Data Explorer)