Generate a forecast fast, without leaving the browser.
Upload a time-series CSV, inspect the data, and generate a practical forecast with confidence bands. This keeps the original tool intention: quick forecasting for ops teams who need answers now.
Best for quick demand or workload forecasting when you already have a historical series and want a fast visual answer.
Private by default
The CSV stays in your browser while the tool parses, fits, and visualizes the forecast.
Forecast first
Upload data, tune the horizon and seasonality, and get an answer without setting up a full WFM system.
Same Soon style, different job
This stays a forecasting tool. Coverage auditing now lives alongside it instead of replacing it.
Step 1: Upload historical data
Bring in the series you want to forecast
Upload a CSV with a date or timestamp column and one numeric series such as tickets, calls, visits, or workload. The tool infers the fields and keeps everything in the browser.
Drop a CSV here or choose a file
Best results come from clean historical series with a steady cadence.
Forecasting guardrails
- Works best when one column represents time and one column represents the metric you want to project.
- At least two full seasonal cycles usually gives a more stable forecast than a very short history.
- Your CSV stays local while the tool parses, visualizes, and computes the forecast.
Wrong question?
If the problem is desk coverage or activity allocation rather than future demand, switch to the coverage audit.
Open coverage auditSeries
History window
Detected cadence
Based on the average spacing in the CSV
Status
Ready to forecast
Adjust the settings and generate the projection
No data available
Upload a CSV file to see visualizations
Trend
Time Series Data
Data Statistics
Mean
Median
Min / Max
Standard Deviation
Forecast Metrics
MAPE (Mean Absolute Percentage Error)
RMSE (Root Mean Square Error)
MAE (Mean Absolute Error)
Outliers Detected
View Outliers
Data Table
Forecast Results
MAPE
Outliers Detected
Future Forecast
Predictions for future time periods (after your input data)
Full Time Series
Historical data with fitted values and future forecast
Forecast with Confidence Intervals
Shows the uncertainty range around the forecast
Step 2: Shape the forecast
Tune the method, horizon, and confidence
Keep the defaults for a fast read, or change the settings when you know the series has a specific cadence or seasonality.
Upload a historical series first. Once a CSV is loaded, the tool will infer cadence and unlock forecast settings.
Observations
Detected cadence
Based on average interval in the CSV
Suggested seasonality
Auto-detected from the historical spacing
Forecast setup
Auto tries Prophet first, then falls back to the browser-side enhanced forecast.
10 periods
This uses the same cadence as the uploaded series. If the data is daily, a horizon of 14 means 14 future days.
Pattern assumptions
Auto suggests 7 periods
Daily data usually uses 7 for weekly patterns. Hourly data often uses 24 for a day-level cycle.
Advanced cleanup
Flag unusual historical points that may distort the result.
Lower values flag more points. A threshold around 3 is a practical default for a first pass.
Switch lenses when needed
Forecast demand here. Audit coverage there.
This page stays focused on future demand or workload. If the problem is who keeps the desk covered and what people are doing inside their shifts, the coverage audit is the better route.
Open coverage audit