FinDB Presentation

Transcript

1. FinDB Group 5: Mihir Arora, John Borkowski, Joshua Dierker, Mabel Egbirika, and Steven Nguyen
2. Meet Our Team Joshua Dierker Full Stack Engineer John Borkowski Backend Engineer Mabel Eqbirika Frontend Engineer Mihir Arora Frontend Engineer Steven Nguyen Backend Engineer
3. About FinDB • Purpose of FinDB is to establish a link between stocks, indexes, and market sectors • Here’s a visual representation of the S&P 500 index… Index: • An assortment of stocks, organized in sectors Sector: • Each stock has a market sector • Many stocks share same sector Stock: • It’s a stock… • attributes like market sector and industry • wowee… finance! C Index Sector A B ^Stocks^
4. Model Relationships Models: • Stock: Has a link to its Sector, as well as links to the top index it is in (Stock-to-top-index). • Index: Has links to top sectors (Index-to-sector), and links to top stocks (Index-to-top-stocks) • Sector: Has links to top stocks in the sector, and indexes with the most weight in that sector. Many-To-Many relationship tables: • Index-to-sector: Table that tracks all sectors within the index and their weight percentages. • Index-to-top-stocks: Table that tracks the top ten stock within the index with weight percentages. • Stock-to-top-index: Table that tracks the indexes with the most weight in the underlying stock
5. Data Collection - Yahoo Finance RESTful API - One-time web-scrape for static sector data (as a cost-saving technique) - `yfinance` and `yahooquery` libraries for batch data collection leveraging API - Indexes - Stocks - Comprehensive Data - Market cap - Historical Data - Sector weights - Top Holdings
6. Backend: Filtering, Sorting, and Searching - Overview - Diverse Search and filtering for user needs - Filtering - Use of SQLAlchemy’s `filter` and `filter_by` methods to build query conditions - Filtering by sector - Sorting - Customizable sorting using `asc` and `desc` methods - Sorting using multiple fields such as Ticker, Name, Current Price - Searching - Full-Text searching across all models - Searching and Filtering Specific Models
7. Postman/ REST API - The endpoints of our API that out frontend consumes with GET requests are organized by stock, index, or sector - These models have GET all endpoints that support filtering, searching, and sorting. - Each model also has a corresponding GET by primary key endpoint - The API also has POST, PUT and DELETE endpoints for creating, updating, and deleting model instances but these are not used by the frontend. - In the backend, the flask route queries the DB with SQLAlchemy, applies any sorts or filters, paginates the result and returns a JSON response
8. The Frontend – Navigation, Splash Page, About Navigation Splash Page Models and Instance Pages
9. The Frontend– Sorting, Searching Sorting (by name, ascending) Searching
10. The Frontend - Filtering Filtering (filter by technology sector)
11. Self Critique - What did we do well? - Communication - Worked well as a team - A visually appealing UI - What we learned? - How to construct association tables for many-to-many relationships in our DB - Data Interaction with React - Flask - GCP Deployment - What we could do better? - Searching across all 3 models - Simplifying filtering applications and resets - What puzzles us? - Deploying frontend and backend simultaneously under one GCP project - DB many-to-many relationships - Why the cost of our website on GCP was so high and how we could fix that issue?
12. HitHub Visualizations Pie Chart UML Diagram
13. HitHub Visualizations Cont. Sortable and Filterable Attributes
14. HitHub Critique - What did they do well? - Visuals and UI - Spotify extension - Cards flip on check - Navigation - What we learned? - Ways to make website more interactive (flipping cards, spotify extension, etc.) - How to make a website visually appealing and intuitive (UI, animations, indicators) - What puzzled us? - Filtering: nothing changes when you click the filtering categories - Top 50 tab - What could be improved? - Implement/improve filtering - Change “Top 50” to “Top 50 Songs” for clarity