π Project Story β Azure AI Crypto Tracker
π Inspiration
The cryptocurrency market moves at lightning speed β prices can change by 10% or more within hours.
Many investors miss out on opportunities or suffer losses because they react after the market moves.
We wanted to build a tool that could predict market changes, notify users instantly, and provide risk insights so they could make decisions in real time.
π What We Learned
- How to use AI anomaly detection to identify unusual price movements.
- How real-time data caching improves application performance.
- The importance of user feedback loops using analytics tools like heatmaps.
- Designing for accessibility with dark mode and intuitive UI flows.
π How We Built It
- Data Pipeline β Collected live crypto data via a custom pipeline and stored it in SQL for analysis.
- AI Models β Trained predictive models in Azure Machine Learning to forecast short-term price volatility.
- Alerts System β Automated Telegram/WhatsApp alerts using Logic Apps.
- Frontend β Built a responsive React dashboard with risk analysis charts and gamification.
- Deployment β Deployed backend as Azure Functions and frontend as a Static Web App.
π§ Challenges Faced
- Data Latency: Solved using Azure Cache for Redis to store frequently accessed data.
- Deployment Errors: Missing environment variables caused Azure Functions to fail, fixed by configuring app settings.
- User Confusion: Microsoft Clarity heatmaps showed users couldnβt find the βSet Alertβ button β redesigned UI placement.
π A Bit of Math
We used a volatility score formula to detect anomalies:
$$ \text{Volatility Score} = \frac{\sigma_{\text{price}}}{\mu_{\text{price}}} \times 100 $$
Where:
- ( \sigma_{\text{price}} ) = standard deviation of recent prices
- ( \mu_{\text{price}} ) = mean of recent prices
High scores trigger anomaly alerts.
Built With
- analytics)
- azure-cache-for-redis-**databases:**-azure-sql-database-**apis:**-custom-crypto-price-feed-api-(no-coingecko)
- azure-functions
- azure-static-web-apps
- built-with-**languages:**-python
- chart.js-**cloud-services:**-azure-machine-learning
- clarity
- javascript
- microsoft
- tailwind-css
- telegram-bot-api
- typescript-**frameworks-&-libraries:**-react.js
- visual-studio-code
- whatsapp-cloud-api-**tools-&-platforms:**-github
Log in or sign up for Devpost to join the conversation.