Live Tweeting

during academic presentations

THE PROBLEM

Presenters at academic conferences often face the challenge of engaging both in-person and virtual audiences simultaneously.

The primary motivation behind this project is to facilitate real-time engagement on social media for both presenters and audience members at academic conferences. Presenters often struggle to connect with their online audiences while delivering their talks, which can hinder effective communication and outreach. Automating live tweeting alleviates this burden, allowing presenters to concentrate on their delivery while the application manages research dissemination. Additionally, audience members can actively participate by sharing insights and engaging in discussions in real time. This dual approach not only enhances the interaction between the physical and virtual attendees but also creates networking opportunities for presenters to connect with fellow researchers. Timely posts significantly influence the visibility and engagement rates of tweets, fostering ongoing conversations around the topics presented and strengthening the overall social media presence for all users involved.

HIGH LEVEL TIMELINE

Literature Review, System Design & Implementation, User Testing and Iteration, Final Evaluation and Refinement

KEY GOALS

Simplicity

Real-time responsiveness

User control and personalization

Minimal cognitive load

UNDERSTANDING THE USER

The project targets both presenters and audience members at academic conferences.

Presenters increasingly rely on platforms like Twitter (now X) to promote their research, as social media can enhance the visibility and long-term reach of their work, providing an alternative channel for discovery that traditional journals may lack. The literature indicates that timely posts are crucial for maximizing engagement, as significant interaction often occurs shortly after a presentation.

For audience members, live tweeting allows for real-time discussions and facilitates engagement with both in-person and virtual attendees, broadening access to information and networking opportunities. It also allowes audience members that are not familiar with the topic be able to tweet about it effortleslly.

However, despite these advantages, active participation in live tweeting is often low, hindered by factors such as demographics, niche topics, and presenter hesitancy.

By understanding these dynamics, the system can be tailored to support both presenters in managing their online presence and audience members in engaging actively during conferences, ultimately enhancing knowledge dissemination and collaboration within the academic community.

Home Page

Serves as the initial entry point, providing a brief overview of the application's functionalities and prompting users to log in or start a presentation.

Onboarding Page

allows users to input optional information about their presentations, such as the title, conference name, and relevant hashtags. Users can also select their preferred tweet generation frequency, which determines how often tweets are generated during the presentation.

Live Tweeting
Page

where users can start, pause, and stop their presentations, with real-time display of generated tweets and a text box for live instructions.
The tweets are generated using a the OpenAI api and persoanlised using the information from the onboarding page, real-time instructions from the user and implicit personalisation that avoids certain tweet styles when a user rejects a proposed tweet.

Tweet History
Page

provides users with a history of all generated tweets, allowing them to view, repost, or manage tweets that were accepted or rejected or even just ignored in previous sessions.

Evaluation & Iteration

The evaluation and iteration process for the live-tweeting system was a critical part of its development, involving three distinct rounds of user evaluations to refine its functionality and user experience.

The first interim evaluation focused on the static design of the live tweeting page, where four participants provided feedback on the layout, clarity of controls, and overall intuitiveness. This stage revealed the need for clearer labeling of tweet approval buttons and improved onboarding guidance.

The second evaluation engaged participants with the fully functional application, allowing them to experience real-time interactions while listening to academic presentations. This hands-on approach led to insights about the intuitiveness of the tweet approval process and the effectiveness of live instruction feature.

Finally, the third evaluation assessed the overall user experience after incorporating feedback from the previous rounds, focusing on usability metrics and the relevance of generated tweets. Iterative changes based on user suggestions enhanced the platform's accessibility and functionality, ensuring it met the needs of both presenters and audience members.

Key Learnings

Importance of User Feedback

Continuous user evaluations provided critical insights that guided design improvements, ensuring the system aligned with user expectations and needs.

Balancing Automation and User Control

Striking a balance between automating tweet generation and allowing users to maintain control over the content was vital for enhancing user confidence and satisfaction.

Clarity in Interaction Design

Specific suggestions to improve the clarity of controls, like renaming the "Start Live Tweeting" button to "Start a Presentation," helped streamline user interactions and reduce misunderstandings during presentations.