This repository contains the frontend and backend for RealmCharts, a live dashboard on RotMG player counts and Steam reviews.
Backend uses Python for data collection, Flask + Gunicorn + nginx for webserver. Frontend is raw HTML/JS + Highcharts.
- Set up Flask app
- Set up automated job to scrape players and online status
- Decouple collection and database logic
- Design frontend basic layout
- Create frontend dashboard for players
- Update player statistics live on dashboard
- Automate playercount removal when server is offline
- Set up automated job to scrape Steam reviews
- Add mobile-friendly frontend
- Create basic frontend for steam positive/negative reviews over time
- Create automated detector when data source is flatlining and auto set to null
- Create more insightful %positive reviews callouts over the current viewed time period, not just entire time span
- Create automated job for populating new table with count from every 5 minutes (so time series forecasting techniques can handle)
- Decide on forecasting model by performing offline analysis
- Deploy basic forecasting model
- Finish About page
- Automated deploying of forecasting model
- Handle maintenance/untrustworthy data periods for forecasting
- Documented API at API Reference
- Error/autocorrelation analysis on prediction errors for model tuning
- Automate new season determination
- Create seasonal callout card comparison (lack of data)
- Replace the cat with season callout box on player count chart
- Describe daily, weekly and RotMG season seaonality analysis
- Statistical analysis of reviews