Moodle and Portal are platforms which students in HKU use a lot. However, the user interfaces of those platforms are not well designed. For example, someone wants to check the deadline of an assignment, he must log in to Moodle, enter his account name and password, find the course among 20 or more courses on the main page, go through half the course page to find the link of that assignment, and finally he can find the due date there. It usually takes 30 or more seconds to do that.
It is only a simple case. According to the data collected from our survey, it was found that Moodle are being used for downloading course materials and submitting assignments, and Portal are being used for course selection, checking timetable and checking course grades in most cases. The wasted time in input and output cumulated is enormous, given that these tasks are being done repetitively during the 4-year university life.
| Tasks on Moodle | Rank |
|---|---|
| Downloading course materials | 1 |
| Submitting assignments | 2 |
| Checking deadlines | 3 |
| Discussing on forums | 4 |
| Watching lesson recordings | 5 |
| Texting tutors and professors | 6 |
| Other | 7 |
In our project, we planned to create an intelligent assistant, which is a chatbot in a form of a web application, to retrieve the information from Moodle and Portal through voice or text commands. For example, when a user says or types “show me my timetable”, the assistant can show the image of the timetable on the screen within 2 to 3 seconds. The user can also create new shortcuts on the main page, so that a button named “show timetable” will return the timetable. It can saves time even further from giving commands. If time allows, more smart features will be developed to make the assistant being more capable, such as predicting user behavior. It can proactively suggest commands to user at the right time.
We will use Vue.js as frontend.
We will use Python as machine learning.
We will use Node.js and Python as backend.
