System Requirements Document (SRD)

Project Title: fierce-dashboard

1. Introduction

The fierce-dashboard project is an advanced Streamlit-based web application designed to predict stock prices using an LSTM model trained on historical stock data. This system aims to provide users with an interactive financial dashboard that combines technical analysis, machine learning predictions, and chatbot support for educational and practical use.

Developed for a global audience, the application supports stock ticker inputs from Yahoo Finance, making it versatile and accessible worldwide. The chatbot is designed to provide both technical explanations and user guidance, ensuring users can navigate the dashboard effectively while learning about financial concepts.

This document outlines the system requirements for the fierce-dashboard, ensuring alignment with the project's objectives and constraints.

2. System Overview

The fierce-dashboard is a modular system that integrates data collection, preprocessing, feature engineering, machine learning, and user interaction into a cohesive platform. The key components include:

1. Data Collection: Fetch historical stock data from Yahoo Finance based on user-selected stock tickers, start dates, and end dates.
2. Data Preprocessing: Handle missing values, normalize data, and split it into training and testing sets.
3. Feature Engineering: Compute technical indicators such as SMA, RSI, and MACD for enhanced analysis.
4. LSTM Model: Train a deep learning model to predict stock prices using historical data.
5. Interactive Dashboard: Provide users with a premium UI for visualizing metrics, charts, and predictions.
6. Chatbot Support: Offer rule-based chatbot assistance for technical explanations and dashboard navigation.

The system is designed to replicate the aesthetics and functionality of modern trading platforms, ensuring a professional and engaging user experience.

3. Functional Requirements

As User:

• I should be able to input any global stock ticker supported by Yahoo Finance.
• I should be able to select a start date and end date for historical data collection.
• I should be able to view metric cards displaying current price, predicted price, percentage change, and RMSE/MAE.
• I should be able to interact with a large, zoom-enabled price chart with hover tooltips.
• I should be able to navigate between tabs for Overview, Technical Indicators, Predictions, and Chatbot.
• I should be able to view charts for SMA, RSI, and MACD with appropriate styling and thresholds.
• I should be able to view historical and forecasted stock prices on a single chart.
• I should be able to interact with a chatbot for technical explanations and guidance on dashboard usage.

As Admin:

• I should be able to monitor the system's performance metrics (e.g., RMSE, MAE).
• I should be able to update the chatbot's rule-based responses for improved user interaction.

4. User Personas

1. General User

• Description: Individuals interested in stock market analysis and predictions.
• Goals:
  • Analyze stock trends and predictions.
  • Learn about financial indicators and machine learning concepts.
  • Use the chatbot for guidance and technical explanations.

2. Admin

• Description: System maintainers responsible for ensuring the dashboard's functionality and accuracy.
• Goals:
  • Monitor system performance.
  • Update chatbot responses and maintain the rule-based logic.

5. Visuals Colors and Theme

The fierce-dashboard will feature a unique, modern color palette inspired by professional trading platforms.

Color Palette:

• primary: #1F2937 (Charcoal Black)
• primary_light: #374151 (Slate Gray)
• secondary: #F97316 (Amber Orange)
• accent: #10B981 (Emerald Green)
• highlight: #EAB308 (Golden Yellow)
• bg: #111827 (Deep Black)
• surface: rgba(31, 41, 55, 0.8) (Semi-transparent Charcoal)
• text: #F9FAFB (White)
• text_muted: #9CA3AF (Muted Gray)
• border: rgba(255, 255, 255, 0.2) (Subtle White)

6. Signature Design Concept

Concept Name: Interactive Financial Galaxy

The landing page of the fierce-dashboard will feature an Interactive Financial Galaxy, where each stock ticker is represented as a glowing star. Users can:

• Hover over stars to see stock names and brief metrics.
• Click a star to zoom into its "orbit," revealing detailed charts and predictions.
• Drag to rotate the galaxy and explore other stock tickers.

Details:

• Animation: Smooth transitions using @react-three/fiber and @react-three/drei. Stars pulsate gently to indicate interactivity.
• Micro-interactions: Hovering highlights connections between related stocks (e.g., sector-based clustering).
• Color Shifts: Stars change color based on stock performance (green for positive, red for negative).
• User Guidance: A chatbot bubble appears at the bottom, offering tips like "Click a star to explore its details!"

This concept ensures the dashboard is visually captivating and highly interactive, making the first impression unforgettable.

7. Interaction Model & Motion Direction

Landing Page: Parallax

• Layered depth via scroll, with stars and galaxy elements translating at different speeds.
• Real content (charts, metrics) scrolls naturally.

Internal Pages: Static

• Focus on clarity and readability for charts, metrics, and chatbot interactions.

8. Non-Functional Requirements

• Performance: Ensure the dashboard loads within 3 seconds for typical datasets.
• Scalability: Support up to 100 simultaneous users without performance degradation.
• Accessibility: Ensure compliance with WCAG 2.1 standards for color contrast and keyboard navigation.
• Security: Prevent unauthorized access to the chatbot logic or model training modules.

9. Tech Stack

Frontend:

• React (for advanced interactivity via @react-three/fiber and @react-three/drei)
• Streamlit (for dashboard rendering)

Backend:

• Python
• FastAPI

Database:

• NoSQL: MongoDB (for chatbot conversation history)

AI Models:

• LSTM (TensorFlow/Keras)

AI Tools:

• Langchain (for chatbot logic)

Local Orchestration:

• Docker

10. Assumptions and Constraints

• Users will have basic knowledge of stock market concepts.
• The system will not include real-time streaming or authentication.
• The chatbot will operate on rule-based logic without external APIs.

11. Glossary

• LSTM: Long Short-Term Memory, a type of recurrent neural network used for time-series predictions.
• SMA: Simple Moving Average, a technical indicator for stock trends.
• RSI: Relative Strength Index, a momentum indicator measuring stock price changes.
• MACD: Moving Average Convergence Divergence, a trend-following indicator.
• OHLCV: Open, High, Low, Close, Volume data for stocks.