Skip to main content Start main content


Q1 How can I start using the UBDA platform?

Before becoming the UBDA user, please check the Charging Model as well as the Services provided by UBDA platform. The 3 simple steps to join the UBDA platform and execute your code are as follows:

  • Step 1: Registration - Register an account to access the UBDA Platform Click here. Login required with your PolyU account and the registration for UBDA account requires at least 2 - 3 business days.
  • Step 2: Check the user guides and related documents. After receiving the account creation confirmation, please check the user guides and related documents in the UBDA usage example documents section.
  • Step 3: Data transfer and code execution
Q2 Who can use the UBDA platform?

UBDA platform is accessible to PolyU researchers, including academic staff and students.

Q3 Is the UBDA platform free of charge for a PolyU staff and student?

Each PI group is assigned a free quota to use the UBDA platform. Annual Max Waived Fee per PI group on computing resources (annual review): 5000 HKD. After the usage of given free quota, charges will be applied based on UBDA charging model.

The Principal Investigator (PI) of the project will be responsible for his/her student's UBDA account and charges. An invoice will be sent to a user based on usage and fund transfer will be arranged by FO colleagues accordingly. Please refer to the Charging Model for more details.

Q4 What are the limitations of a user, in terms of storage space and cpu/gpu allocations?

The limitations depend on the “UBDA Computing Intensive Service” and “UBDA Big Data Service”. For “UBDA Computing Intensive Service”, there are limitations on “Job Queues' Resource Limit” and “User Storage Quota” (referring to Services). For the VM mode of “UBDA Big Data Service”, there are configurations as referring to Charging Model.

For the different demands from different projects, a user should let us know what computing resources they need, like no. of CPU cores, GPU, storage size as well as the applications to setupon the UBDA platform. We are pleased to work with you to setup and build for running your desired programs/application. For such requirements, users should let us know their plans to reserve the resources.

Q5 What are the UBDA services available?

UBDA provides a dedicated, secure, and scalable 24/7 big data platform for building the analytic solutions.

Q6 What are the advantages of joining UBDA platform as opposed to starting my own?

UBDA is a big data platform to store and analyze your data for finding the hidden patterns, exploring unknown correlation, improving prediction, supporting decision making, recommending services, and products and other analytic solutions.

UBDA offers several benefits, including:

  • GPUs and CPUs accelerated computing platform
  • Consultancy platform to formulate the big data related research problems
  • Opportunity to have joint labs, projects, and sponsorships
Q7 There are 3 UBDA systems. How I choose which one to use?

The 3 UBDA systems are the “(A) UBDA Computing Intensive Service”, “(B) UBDA Big Data Service – Virtual machine” and “(C) UBDA Big Data Service – Online JupyterLab Development Tool with GPU”.

  1. It is a batch system which a rich set of job scheduling and workload management. Each job can assign or allocate different resources (CPU, GPU, memory, disk storage, etc.) during its execution. Also, this system is relatively less cost in term quotas.
  2. It is a virtual machine service. A virtual machine is a logical representation of a physical computer which can deploy one of different system configurations available. For the user, the VM is fully dedicated to 24x7 services solely.
  3. It is Jupyter Notebook with a web-based interface for Python program development. Some advantages are using Jupyter Notebook as referring and some examples can be found in UBDA web site in Python examples for ML/DL and RAPIDS Examples

Q1 I have a research project which would like to apply AI and machine learning techniques. Any UBDA staff advice on modelling and/or algorithms?

UBDA offers several benefits, including consultancy platform to formulate the big data related research problems and opportunity to have joint labs, projects, and sponsorships.

UBDA staff have good experiences of computing and applied mathematics works. If your project is associated with UBDA, we can join your application/research team as a consultant for providing advice on modelling problems and/or designing algorithms.

Q2 I am developing/writing a research proposal of applying AI and machine learning techniques. Any UBDA staff advice on the adoption of the techniques and required resources related to UBDA?

UBDA staff can help prepare the research proposal by providing techniques and required resources related to UBDA. We can also help you in analyzing your problem and formulating the data-driven research issues.

If needed, we can join your application/research team.

Q3 I am interested to apply AI and machine learning techniques for a research project. Are there available use cases and/or tutorial references?
Each UBDA account provides some Python examples of ML/DL, such as classification, clustering, text analysis, convex optimization, etc. The details to run these examples are given in the Documents
You can run these examples to get familiar with the UBDA platform.
Q4 Is UBDA support data sharing among researchers working on the same project?
You can send us a separate request. We will create a group for sharing the data. Users will run their code/application within their UBDA account.

Q1 My job is spending a lot of time in the queue. Why? When will it start?

You need to queue when your request resources are using by other users. Your job will start when your requested resources are available

Q2 How can I reduce the amount of time my job spends waiting in the queue?

Due to the large requested resources may need a lot of time to waiting resources. You may to reduce the amount of waiting time.

Q3 My batch job finishes as soon as there is no output. Any problem?
An output file saved when the job finish. You may find the output file in directory where your PBS stored.
Q4 Can I transfer files from own local dataset to the UBDA platform?
Yes. You can download the guideline at GUI Files Transfer (Mac iOS)/GUI Files Transfer (Windows System)

Q1 What are the different VM configurations/images? How I choose which configuration should use?

We have different VM configurations. You can visit our Charging Model to know more.

Q1 What is the UBDA Jupyter configurations?

The UBDA Jupyter supports Python version 3.x with most of the packages. Additional packages can be installed by entering ‘!pip install <packageName>’ in a prompt line inside the workspace. You can visit Charging Model to know more.

Q2 Is the cost of using the UBDA Jupyter services?
Yes, please refer the charging model web page for the section of "JupyterHub Service"

Q1 What are sample programs/manuals available?

You can visit Documents

Q2 How UBDA will support the update of a software's version?
You can send us a separate request. Applications will be reviewed regularly and will be applied once it is well-tested and integrated with the UBDA platform smoothly.

Q1 What is the procedure to install SAS and Stata on UBDA clusters, if we have a license for the same?

Users need to present license for running their license-required applications in UBDA platform. We are pleased to work with you to test your application with your provided legal license.


Will UBDA provide any dataset for the data analytics? Any public data sources?

No, UBDA will not provide any dataset for your research purpose. On the other hand, UBDA staff can suggest public and accessible dataset available. For example, Google Cloud Public Datasets (, Kaggle (, and Hong Kong Open Government Data(



Your browser is not the latest version. If you continue to browse our website, Some pages may not function properly.

You are recommended to upgrade to a newer version or switch to a different browser. A list of the web browsers that we support can be found here