Shenzhen Web Science Summer School 2017

Shenzhen Web Science Summer School 2017
The Shenzhen Web Science Summer School was held between 20 and 24 March at the Tsinghua-Southampton Web Science Laboratory, Shenzhen, China. PhD students from Tsinghua, and the Web Science Institute, University of Southampton, UK, worked together on two Data Challenges in a competitive datathon. The aim was to learn and put into practice new data analytic skills which students’ could apply to their own research. From this collaborative work, innovative visualisations and code were created and shared on the Web Observatory.

The students were split into two groups, and provided with 20 million text files from Chinese newspapers and 5 million text files from UK newspapers. Each group chose a Data Challenge to work on.  Group 1 tackled the Disaster Management Challenge and Group 2 worked on the Shared Bicycle Scheme Challenge.

As well as the data they were provided with, both groups needed to search for and explore any additional, relevant data that could help them give further context to their investigations. This proved difficult, as Web sources that are widely used by UK-based students were not available in China (no Google!!). This, along with some cross-cultural and language misunderstandings, added to the challenge. However, hurdles were overcome and the teams worked together to produce insightful analyses.

Group 1: Disaster Management

Group 1 - Tsinghua Summer School
Group 1

Tsinghua University: Wang Chen, Jinxin Han, Xin He, GengBiao Shen, Kan Wu, Jing Zhang
WSI: Jo Munson and Sami Kanza
Mentor: Eugene Siow
This Challenge explored patterns of flooding in the real world and how it is reported on social media. Aided by their mentors expertise in Statistical Modelling, Behavioural Mining and Sentiment Analysis, students were helped to evaluate the propagation of online discussion on flooding. By applying  a Natural Language Processing algorithm the team developed a prototype dual-language dashboard which mapped social responses to flooding.
Outputs: Main websitePresentation Slides

Group 2: Shared bicycle schemes in China

Ofo and Mobike bicycles at Dongdan/N509FZ ©2017/cc-by-sa-4.0
Ofo and Mobike bicycles at Dongdan/N509FZ ©2017/cc-by-sa-4.0

Tsinghua University: Wei, Haimei, Jiamei, Shuo
WSI: Bart Paszcza and Chira Tochia
Mentor: Xin Wang
In recent years bicycle sharing projects have sprung up around the globe – helping to solve the ‘last mile’ problem, and enabling people to quickly and easily travel to and from major transport hubs. In the US there are more than 120 shared bicycle projects covering millions of miles every month, while in China shared bicycle projects, Mobike, and Ofo  are also attracted a great deal of interest. For the datathon the team developed a number of data visualisations showing use of bike sharing schemes in Shenzhen.
Outputs: HeatmapClustermapBike journeysPresentation Slides

Group one's Flood visualisation
Group one’s Flood visualisation

And the winners were…group one! A very well deserved win for such beautiful and useful visualisations for mapping floods in both Chinese and English.

Leave a Reply

Your email address will not be published. Required fields are marked *