Librari: Intelligent Digital Library

Inspiration

Our inspiration for Librari stems from the frustration of feeling lost in the book selection process after finishing one. Like the need to unify (Segment Unify) all aspects of reading journey, Librari aim to simplify the way you discover your next favorite book.

What it does

Librari is like your smart reading companion. It uses advanced tech to suggest books you'll love, analyzes what you read, and keeps things easy with a user-friendly setup. It's your go-to for discovering great reads in a massive digital library. It comes with a recommendation section that helps you to find the next book based on your preferences.

How we built it

Librari is designed with simplicity and efficiency in mind. For the frontend, its using Next.js. On the backend, I rely on Flask to fetch predictions from our recommendation model. Then ALS algorithm (collaborative filtering) to build intelligent book recommendations, and leveraged Twilio Segment and Stitch data to ensure a complete view of user's reading journey, while BigQuery served as the backbone for efficient data storage and retrieval.

Challenges we ran into

Building Librari came with its share of challenges. Deploying the recommendation model was a novel experience, and finding the right dataset proved to be quite a task. Daily progress was slower than anticipated, especially since I was managing everything solo while juggling a busy schedule.

Accomplishments that we're proud of

I'm extremely proud of the knowledge I've gained in a short amount of time, particularly in areas like Apache Spark, Twilio Segment, and BigQuery. These were completely new to me, and diving into analytics with Twilio Segment opened up a plethora of fresh ideas for software development. Despite facing challenges and some gaps in the software, I'm proud to say that I successfully completed the project, showcasing my ability to learn and implement new technologies.

What we learned

When creating Librari, I got hands-on experience with the ALS algorithm for smart recommendations. And I learned how to deploy machine learning models and optimize data handling using BigQuery. Figuring out how analytics and machine learning work together, especially with Twilio Segment for data collection, was eye-opening.

What's next for Librari: Intelligent Digital Library

I'll be enhancing personalized recommendations by grouping users with similar reading interests. Expect features like interactive discussions and reviews for authors. I'll also be adding sentiment analysis to capture the emotional tone of content. And guess what? It's not just about books – I'll be expanding to include digital assets like movies and music

Built With

• apache-spark
• dataproc
• flask
• google-bigquery
• google-cloud
• next.js
• segment
• stitch