Inspiration
We wanted to understand how chronologically ordered events relate to the trends in the posted dataset from state-controlled media sites, and to test the validity of the Chinese meddling in Canadian elections.
What it does
The data provides insights into the activities of Chinese News Media Outlets such that we could discover the possible events and reasons for spikes in activity to understand foreign influence from China to Canada.
How we built it
- Used Google's YouTube Data API to gather all the video data on channels that existed on the spreadsheet
- Graphed this information against time to track how these organizations posted relative to current events
Challenges we ran into
- Finding APIs that gave us an adequate amount of data
- The Twitter/X API was too limited in its capacity to gather large amounts of data
- Facebook API was very difficult to use and getting the academic permission to scrape large amounts of data was hard
- YouTube API use was successful, but involved creating several different API keys/accounts to avoid the quotas on the free use of the API
- Formatting Graphs to be readable while including necessary information
- Graphs could not be too big, but also needed to include the variety of different data we wanted
Accomplishments that we're proud of
- Created graphs of the posted datasets and were able to break it down by language, region of focus, and parent entity
- Allowed us to analyse the effects these variables had on when news accounts posted
What we learned
- China had a spike in media output in late 2015 (November - December) most likely due to the first ever diplomatic meeting between the Chinese and Taiwanese heads of state in decades.
- The graphs of the organizations controlled by The Ministry of Foreign Affairs, Shanghai Media Group, Chinese Media Group, and Xinhua News Agency all had noticeable increases in posts after April 2020 (COVID-19 Pandemic).
- The Facebook accounts of the Chinese Media Outlets have the most traffic, and should hence be monitored in order to combat the Chinese narrative.
What's next for CANIS_U of T
- Getting more post data from other social media sites (Twitter/X, Facebook, TikTok, Reddit, etc.)
- Understanding the reach these videos have and where these channels were reaching the most people
- Most importantly, identifying when/where Canadians are viewing this content from and understanding how that affects voting data
Analysis
Analysis are all included in the presentation slide
Log in or sign up for Devpost to join the conversation.