HCI pleased to announce it’s annual summer training to FCIH students, We are seeking highly motivated students to join our lab and groups. The training starts at 22 June 2014 and stands till end of September 2014.

https://docs.google.com/a/fcih.net/spreadsheet/viewform?formkey=dHZSdC1aR2NlRF95czBBWHFRVDM0ZHc6MA

This year we are going to hold a group project divided as follows

What is data visualization? Why is it needed? How can it contribute in understanding the data and help
in choosing the right model that ts this data? Data visualization, from its name, is about presenting and
displaying data in one or more informative way to either exploring to help understanding data and its
nature, or presenting data to convey to audience particular messages.
For more information on data visualization and its methods, refer to Chen et al. (2008); Chapter 2 of
the book accounts as a good review of the eld.
2 Project Description

The objective of this small project is to implement some standard data visualization methods. The methods
implemented should be beautiful, as well as, scientic and informative.
For the scientic aspect, not all data visualization techniques are simple or straightforward like scatter
plot. There are some techniques that require some mathematical treatments before plotting, e.g., Multi
Dimensional Scaling (MDS).
For the aesthetic aspect, you will need to learn web technologies, e.g., HTML5, CSS, JS, and D3 library
(Bostock et al., 2011). Please, visit http://d3js.org/ for a very nice pool of methods implemented in a
beautiful way.

At the end of your training you should have implemented some of those techniques and contributed to
the D3 open source community.

References
Bostock, M., Ogievetsky, V., Heer, J., Dec. 2011. D: Data-Driven Documents. IEEE transactions on visualization
and computer graphics 17 (12), 2301{9.
URL  http://www.ncbi.nlm.nih.gov/pubmed/22034350
Chen, C.-h., Hardle, W., Unwin, A., 2008. Handbook of data visualization. Springer, Berlin.
URL http://www.loc.gov/catdir/toc/fy0804/2007931825.html

The project will utilize the FTIR and Large display projection to build a seamless interaction board that can allow (transferring and sharing of data) across different platforms and output/input devices. The project will aim to build a multimodal input environment with gestures, leg,gaze tracking….etc.

Please read and summarize the vacuum-chi2005-bezerianos or laser pointer interaction with large display.

Please write a code that can allow TCP/IP client server communication to send some text strings.

This project aims at implementing several dimensionality reduction techniques to help in visualizing high-dimensional data in two or three dimensions (ref: attached file). Moreover, similar techniques could be used for large graph drawing and implementing several graph analytics, such as community detection. The focus will be on heavily optimized implementation. In particular, there will be two implementations: one in C++ and one on GPU. Both variants will rely on high-level optimizations (algorithm and data structure), but the C++ variant will also rely on low-level optimizations (CPU cache management, prefetching, vectorization/SIMD, etc.)

Write 1-page summary of the Tenenbaum (the paper that introduced Isomap, a famous dimensionality reduction technique).

If you choose the C++ track, write a C++ program that uses the SSE compiler intrinsics in order to perform scalar product (inner product) of two vectors of dimension 4. If you choose the GPU track, write the same program using OpenCL. For both tracks: Preferably use GCC on Linux, but MSVC on Windows is also acceptable.