In September 2020, HKSYU launched three laboratories for the study of big data, virtual reality and robotics. The intent is to reinvent liberal arts education, enabling the university to stay relevant in the digital age. By adopting digital technology and online teaching tools, HKSYU will make learning easier and more interesting.

One lab, the Big Data Laboratory, was designed to facilitate studies and support research projects requiring data analytics, machine learning and data visualization. The challenge, however, was to equip the lab with technology that both students and faculty could easily learn and use.

Dr. Connie Yuen, Head of the Department of Applied Data Science, Director of iFREE GROUP Innovation and Research Centre, and Director of the Big Data Laboratory, was tasked with solving the problem. “The students are interested in learning more about technology and how to use it in their studies and lives,” she explains. “But they don’t know how. It’s the same for the faculty members. They would like to use tools to support their research and data analysis, but don’t know where to start. It’s complicated for them.”

Dr. Yuen chose IBM Cloud Pak for Data technology, which runs on the university’s on-premises private cloud. “Cloud Pak for Data makes learning easier for students and teachers who are starting their journey with data analysis,” adds Dr. Yuen. “That’s why it was the right choice for us.”

At the start of the engagement, the IBM Client Engineering team conducted a workshop to understand the university’s pain points, expectations and requirements for academic use cases. The team also learned how HKSYU wanted to take advantage of machine learning — and natural language processing in particular — to uncover insights from unstructured data, determine the best topics for research and deliver high quality analysis within a reasonable timeframe.

The workshop resulted in a two-week minimum viable product (MVP) of IBM Watson® Discovery software for news analyses. HKSYU selected the data sources. The results proved the technology’s ability to ingest and analyze news data. The university was impressed with the software’s comprehensive user interface to help students extract meaningful insight, verify the correlation of topics and keywords, and generate reports.

The platform also provides an array of analytic components, but gives the university the flexibility to select just those services and components it needs most. One of those components is the AutoAI graphical tool in IBM Watson® Studio technology, which is integrated with the IBM Cloud Pak for Data platform. The tool automatically runs the key tasks used to build machine learning models, such as data pre-processing and model selection. Plus no coding is needed.

“With AutoAI, students can visualize the whole modeling process, starting from data collection to data analysis to results and the performance of the algorithms,” says Dr. Yuen. “It’s very easy to use. Within a three-hour class, the students learn from the demo, listen to a lecture and practice in a two-hour lab. The professors are also more comfortable teaching data analysis because it’s easier.”

One professor also used IBM Cloud Pak for Data to teach fintech students the Python programming language. “Because students can easily visualize the data, it motivated them to learn Python,” adds Dr. Yuen. “They also have all the tools they need right there to do data clean up, so the task is easier.”