Inspiration
We recognized that parliament hearings can be lengthy and difficult to comprehend, yet they have a significant impact on our lives. We were inspired by the concept of a TL:DR (too long, didn't read) summary commonly seen in news articles, but not available for parliament proceedings. As news often only covers mainstream headlines, we saw the need for an all-inclusive, easy-to-access summary of proceedings, which is what led us to create co:ngress.
What It Does
Co:ngress is an AI-powered app that summarizes the proceedings of the House of Commons of Canada for the convenience of the users. The app utilizes Cohere API for article summarization, which processes the scraped proceedings and summarizes them into a shorter, more easily understandable version. Furthermore, co:ngress fetches an AI-generated image using the Wombo API that loosely relates to the hearing's topic, adding an entertainment factor to the user experience. For scalability, we designed the app to store summaries and images in a database to minimize scraping and processing for repeat requests.
How We Built It
The co:ngress app is built using a four-step process. First, a web scraper is used to scrape the transcripts of the parliament hearings from the House of Commons website. Next, the scraped article is passed to the Cohere API for summarization. After that, the topic of the parliament proceeding is interpreted, and the Wombo API generates an image that relates to it. Finally, the image and summary are returned to the user for easy consumption. The app was built using various tools and technologies, including web scraping, API integration, and database management.
Challenges We Ran Into
One of the biggest challenges we faced was fine-tuning the Cohere API hyperparameters to achieve optimal results for our app. This process took up a lot of time and required a lot of trial and error. Additionally, we had to ensure that the web scraper was able to efficiently gather random samples from the parliament website without errors.
Accomplishments That We're Proud of
We are proud of the modular approach we took in building our app and how we were able to connect all the different pieces together to create a seamless user experience. Additionally, we were able to successfully summarize parliamentary proceedings into a useful result that can be easily understood by the average Canadian.
What We Learned
One of the key things we learned was the power and versatility of the Cohere API, which has impressive training data that helped us create accurate summaries. We also learned the importance of taking a modular approach to app development, which can make the process more efficient and less prone to errors.
What's Next for Co:ngress
For Co:ngress, the next steps would be to implement user feedback and continually improve the summarization algorithm to ensure the accuracy and relevance of the summaries. The team could also explore expanding the app to cover other areas of government proceedings or even other countries. Finally, as mentioned, creating a scalable cloud infrastructure to support a large number of users would be an important next step for the app's growth and success.
Log in or sign up for Devpost to join the conversation.