Sahil MLflow studio
Market Snapshot
Follow live BTC price movement, ensemble forecasts, model health, and prediction quality in a refined interface built for desktop and mobile.
Runtime State
Active
Refresh Cycle
5 Minutes
Focus
BTCUSD Hourly
Studio Online
Current BTC Price
$ --,---.--
Latest cached close pulled from local market data.
Realtime view of price, forecasts, and platform status.
Realtime Forecast Brief
Read the market pulse before the next hourly close.
This spotlight area pairs the live BTC reference price with the latest ensemble projection so the core signal stands out immediately.
Current Price
$ --,---.--
Next Hour Forecast
$ --,---.--
Target Time
--
Quick Signal
Realtime status panel
Inference cadence
On demand
Ensemble accuracy
94.2%
Model health
Optimal
Current BTC Price
$ --,---.--
Current market reference pulled from the local cached OHLC stream.
Realtime feed available
Next Hour Forecast
$ --,---.--
Latest ensemble output produced by the active forecasting stack.
Generating...
Inference Stability
Moderate
Temporal smoothing keeps short-term movement realistic between runs.
Smoothed signal
Operations Status
Nominal
UI, logs, and prediction endpoints are aligned for dashboard display.
All systems green
Forecast View
Prediction trend against actual price
Compare actual closes with ensemble, LR, ARIMA, and LSTM outputs across recent cycles.
Resolved Logs
Realtime prediction feed
Recent entries displayed in a compact operator table.
| Timestamp (IST) | Predicted Time | Actual | Ensemble | LR | ARIMA | LSTM |
|---|---|---|---|---|---|---|
| Loading logs... | ||||||
Top Performing Model
--
Lowest RMSE
--
Status
Optimized
Mean Squared Error
--
Primary ensemble MSE from resolved historical predictions.
RMSE
--
Lower is better and drives the ranking view.
Prediction Stability
Moderate
Recent ensemble consistency summarized into a volatility posture.
Calculating...
Directional Accuracy
--
Measures whether price direction was anticipated correctly.
Leaderboard
Model ranking by RMSE
Fast comparison across the trained models used by the pipeline.
| Rank | Model Name | RMSE |
|---|---|---|
| Loading ranking... | ||
Chart Lens
RMSE comparison
Bar-chart view of error spread across each model.
Clear evaluation flow
Performance metrics surface the strongest model quickly while keeping the broader stack visible for comparison.
Readable analytics
Dark cards and brighter chart accents make it easier to scan between leaderboard and metrics.
Operational focus
Stability, RMSE, and directional accuracy now read as one connected performance story.
Data Drift Status
No Drift
Most recent drift check indicates source data remains within normal behavior ranges.
P-Value: 0.42
Model Drift
Minor Decay
Recent performance suggests mild degradation that may justify a retraining pass.
Retraining recommended
Observation Mode
Resolved Logs
The monitoring chart reads absolute prediction error from completed history entries.
Watch Status
Tracked
Frontend checks the latest monitoring data whenever this page is opened.
Watchtower
Prediction error over time
Absolute error trend across resolved predictions to help spot drift or instability.
Live diagnostics
This area now reads as a dedicated diagnostic view rather than a spare chart dropped into the layout.
Readable risk cues
Warnings and healthy states are color-coded for faster review on desktop and mobile.
Operations
Prediction history
Recent predictions with status, timestamps, and model outputs in a touch-friendly table.
| Timestamp (IST) | Predicted Time | Ensemble Pred | LSTM | ARIMA | LR | Status |
|---|---|---|---|---|---|---|
| Loading history... | ||||||
Inference Latency
--ms
Tracked from inference request to response for the latest manual execution.
API Status
Operational
Frontend and API routes are currently reachable for dashboard rendering.
Operator Note
This view stays mobile-friendly by keeping tables horizontally scrollable instead of collapsing them into unreadable blocks.