PowerForecast Docs
Back to Wizard

PowerForecast Wizard User Guide

This guide walks you through every screen of the Forecasting Wizard with detailed explanations, tips, and troubleshooting.

Overview

The Forecasting Wizard guides you through 5 steps to create and run a forecast:

```

┌─────────────────────────────────────────────────────────────────────────┐

│ Step 1 Step 2 Step 3 Step 4 Step 5 │

│ Business Time Data Model Review │

│ Goal Config Source Selection & Run │

└─────────────────────────────────────────────────────────────────────────┘

```
Step Title Purpose Required?
1 Business Goal Select what you want to forecast ✅ Yes
2 Time Configuration Set frequency, horizon, confidence ✅ Yes
3 Data Source Upload your time series data ✅ Yes
4 Model Selection Choose Auto or Manual models ✅ Yes
5 Review & Run Validate and execute forecast ✅ Yes

Getting Started

Main Dashboard

[Screenshot: Main Dashboard]

The dashboard shows:

Element Description
**Open Wizard** button Start a new forecast or continue editing
**Specification Summary** panel Shows your current forecast configuration
**Recent Runs** panel Displays your forecast history

Starting the Wizard

Click "Open Wizard" to begin creating a new forecast specification.

Step 1: Business Goal

[Screenshot: Step 1 Modal]

Purpose

Tell PowerForecast what you want to predict. This helps the system understand your business context and select appropriate models.

Business Target Options

Target Icon Best For Example
**Demand** 📦 Product demand, sales volume, order quantities "How many units will we sell next month?"
**Workload** 👥 Staff requirements, task volume, capacity needs "How many staff do we need next quarter?"
**Revenue** 💰 Sales revenue, income projections "What will our revenue be next year?"
**Claims** 📋 Insurance claims, requests, incidents "How many claims will we receive?"
**Transactions** 🔄 Transaction volume, payments "How many transactions per day?"
**Custom** ⚙️ Define your own target "Patient admissions", "Energy consumption"

Selecting a Target

1. Click on one of the target cards

2. The selected card will highlight

3. For Custom, enter your target name in the text field that appears

Granularity (Optional)

Granularity allows you to segment forecasts:

Option Use For
No segmentation Aggregate forecast across all data
Per Product Forecast each product separately
Per Region Forecast by geographic region
Per Customer Segment Forecast by customer type
Per Branch/Location Forecast by location
Per Category Forecast by product category

Example: If you select "Per Product" and upload data with multiple products, PowerForecast will create separate forecasts for each product.

Buttons

Button Action
**Cancel** Close wizard without saving
**Next** Proceed to Step 2

Tips

  • Choose the target that best matches your business question
  • Granularity is optional - you can always aggregate results later
  • Custom targets work just like predefined ones

    • Common Issues

    Issue Cause Solution
    "Please select a business target" warning No target selected Click on a target card
    Custom target field not appearing Custom card not clicked Click the Custom card first

    Step 2: Time Configuration

    [Screenshot: Step 2 Modal]

    Purpose

    Define the time structure for your forecast: how often data points occur, how far ahead to predict, and confidence level.

    Forecast Frequency

    Frequency Code When to Use Example
    **Hourly** H High-frequency monitoring Energy consumption, website traffic
    **Daily** D Daily operations Daily sales, daily demand
    **Weekly** W Weekly planning Weekly reports, weekly capacity
    **Monthly** M Monthly planning Monthly revenue, monthly demand
    **Quarterly** Q Quarterly reviews Quarterly sales, quarterly budgets
    **Yearly** Y Long-term planning Annual forecasts, strategic planning

    Default: Monthly (M)

    Forecast Horizon

    How many periods ahead to forecast.

    Horizon Typical Use
    1-6 periods Short-term planning
    6-12 periods Medium-term planning (default: 12)
    12-24 periods Long-term planning
    24+ periods Strategic planning

    Default: 12 periods

    Note: The horizon is measured in the selected frequency. For example:

  • Monthly frequency, horizon 12 = 12 months ahead
  • Daily frequency, horizon 30 = 30 days ahead

    • Confidence Level

    The probability that the actual value will fall within the prediction interval.

    Level Use For
    **80%** Less conservative, narrower intervals When you need tighter ranges
    **90%** Standard business planning Common choice for operations
    **95%** Default - balanced confidence Most business scenarios
    **99%** High-stakes decisions Risk-averse scenarios, compliance

    Default: 95%

    Understanding Confidence Intervals:

  • 95% confidence means there's a 95% chance the actual value will be between the lower and upper bounds
  • Higher confidence = wider intervals (more uncertainty)
  • Lower confidence = narrower intervals (less uncertainty)

    • Scenario (Optional)

    Select a scenario type for your forecast:

    Scenario Icon Description When to Use
    **Baseline** 📊 Most likely outcome Standard forecasting (default)
    **Optimistic** 📈 Best-case scenario Growth planning, best-case analysis
    **Pessimistic** 📉 Worst-case scenario Risk planning, worst-case analysis
    **Stress** ⚠️ Extreme conditions Stress testing, contingency planning

    Default: Baseline

    Note: Scenarios affect how models interpret trends and patterns. The same data may produce different forecasts under different scenarios.

    Buttons

    Button Action
    **Previous** Go back to Step 1
    **Next** Proceed to Step 3

    Tips

  • Match frequency to your data's natural period (daily data → daily frequency)
  • Horizon should align with your planning needs (don't forecast too far ahead)
  • Higher confidence levels provide more conservative forecasts
  • Scenarios help explore different business outcomes

    • Common Issues

    Issue Cause Solution
    Invalid horizon value Number too large or negative Use values between 1-365
    Frequency mismatch Frequency doesn't match data Ensure data spacing matches frequency

    Step 3: Data Source

    [Screenshot: Step 3 Modal]

    Purpose

    Upload your time series data. PowerForecast accepts CSV or JSON files with historical data.

    Data Format Requirements

    CSV Format

    ```csv

    unique_id,ds,y

    product_1,2024-01-01,100

    product_1,2024-02-01,120

    product_1,2024-03-01,105

    product_2,2024-01-01,50

    product_2,2024-02-01,55

    product_2,2024-03-01,52

    ```

    Required Columns:

  • `unique_id`: Series identifier (e.g., product name, location ID)
  • `ds`: Date in ISO format (YYYY-MM-DD)
  • `y`: Numeric value to forecast

    • JSON Format

    ```json

    [

    {

    "unique_id": "product_1",

    "ds": ["2024-01-01", "2024-02-01", "2024-03-01"],

    "y": [100, 120, 105]

    },

    {

    "unique_id": "product_2",

    "ds": ["2024-01-01", "2024-02-01", "2024-03-01"],

    "y": [50, 55, 52]

    }

    ]

    ```

    Required Fields:

  • `unique_id`: String identifier for each series
  • `ds`: Array of date strings (ISO format)
  • `y`: Array of numeric values

    • Uploading Data

    Method 1: Drag and Drop

    1. Drag your CSV or JSON file onto the upload zone

    2. File will be automatically parsed and validated

    3. Preview will appear if successful

    Method 2: Click to Browse

    1. Click the upload zone

    2. Select your file from the file picker

    3. File will be automatically parsed

    Data Preview

    After uploading, you'll see:

    Metric Description
    **Time Series Count** Number of unique series in your data
    **Total Data Points** Total number of observations
    **Date Range** Earliest and latest dates

    A preview table shows sample rows from your data.

    Data Quality Options

    Option Description Default
    **Handle Missing Values** Automatically interpolate missing data points ✅ Enabled
    **Handle Outliers** Detect and handle outliers using IQR method ✅ Enabled

    Missing Values:

  • PowerForecast uses interpolation to fill gaps
  • Missing values are estimated based on surrounding data
  • Critical for maintaining time series continuity

    • Outliers:

  • Outliers are detected using the Interquartile Range (IQR) method
  • Outliers are adjusted to prevent them from skewing forecasts
  • Useful for handling data entry errors or one-time events

    • Buttons

    Button Action
    **Previous** Go back to Step 2
    **Next** Proceed to Step 4

    Tips

  • Ensure dates are in ISO format (YYYY-MM-DD)
  • Data should be sorted chronologically
  • Minimum data points: At least 2× the seasonal period (e.g., 24 months for monthly data with yearly seasonality)
  • More data generally leads to better forecasts
  • Check the preview to verify data was parsed correctly

    • Common Issues

    Issue Cause Solution
    Upload fails Wrong file format Use CSV or JSON with required columns/fields
    "Invalid date format" Dates not in ISO format Convert dates to YYYY-MM-DD format
    "Missing required field" CSV missing columns or JSON missing fields Ensure unique_id, ds, and y are present
    Preview shows no data Data parsing error Check file format matches examples
    "Insufficient data" Too few data points Provide at least 2× seasonal period of data

    Step 4: Model Selection

    [Screenshot: Step 4 Modal]

    Purpose

    Choose how PowerForecast should select forecasting models. You can let it automatically choose the best models, or manually select specific ones.

    Selection Modes

    Automatic Mode 🤖

    What it does:

  • PowerForecast analyzes your data characteristics
  • Performs statistical profiling (trend, seasonality, stationarity)
  • Automatically selects the best model(s) based on data patterns
  • Considers multiple models and chooses the most suitable

    • When to use:

  • ✅ You're not sure which model to use
  • ✅ You want the best model for your data
  • ✅ You're new to forecasting
  • ✅ Your data patterns may change over time

    • How it works:

    1. Data profiling identifies characteristics (trend, seasonality, volatility)

    2. Model selector evaluates which models fit best

    3. Best model(s) are selected automatically

    4. Forecast runs with optimal configuration

    Manual Mode 🎯

    What it does:

  • You choose specific models to run
  • PowerForecast runs all selected models
  • You can compare results from different models

    • When to use:

  • ✅ You know which models work well for your data
  • ✅ You want to compare multiple models
  • ✅ You have specific model requirements
  • ✅ You're experimenting with different approaches

    • Available Models

    Model Best For Characteristics
    **AutoARIMA** Automatic ARIMA selection Handles trends and seasonality automatically, robust
    **ETS** Exponential Smoothing Handles various patterns, good for seasonal data
    **SeasonalNaive** Simple seasonal patterns Baseline model, good for comparison

    AutoARIMA:

  • Automatically finds optimal ARIMA parameters
  • Handles both trend and seasonality
  • Good for most business time series
  • Recommended for beginners

    • ETS (Exponential Smoothing):

  • Flexible smoothing methods
  • Handles additive and multiplicative seasonality
  • Good for data with clear patterns
  • Fast and interpretable

    • SeasonalNaive:

  • Simple baseline model
  • Uses last season's value as forecast
  • Good for comparison with other models
  • Fast but less sophisticated

    • Manual Model Selection

    If you choose Manual mode:

    1. Check boxes next to models you want to use

    2. You can select multiple models

    3. PowerForecast will run all selected models

    4. Results will show which model performed best

    Buttons

    Button Action
    **Previous** Go back to Step 3
    **Next** Proceed to Step 5

    Tips

  • **Start with Automatic** - Let PowerForecast choose the best model
  • **Use Manual for comparison** - Run multiple models to see which performs best
  • **AutoARIMA is versatile** - Works well for most business scenarios
  • **SeasonalNaive as baseline** - Include it to compare other models against a simple baseline

    • Common Issues

    Issue Cause Solution
    No models available Data format issue Go back to Step 3 and verify data
    Model selection fails Insufficient data Ensure you have enough historical data points

    Step 5: Review & Run

    [Screenshot: Step 5 Modal]

    Purpose

    Final review of your forecast specification. Validate everything before running the forecast.

    Specification Summary

    The review screen shows:

    Section Information
    **Business Target** What you're forecasting
    **Frequency** Data frequency (Daily, Monthly, etc.)
    **Horizon** How many periods ahead
    **Confidence Level** Prediction interval confidence
    **Scenario** Baseline, Optimistic, etc.
    **Data Summary** Number of series, data points, date range
    **Model Selection** Auto or Manual with selected models
    **Data Quality** Missing value and outlier handling settings

    Validation

    Before running, PowerForecast checks:

  • ✅ Business target is selected
  • ✅ Time configuration is valid
  • ✅ Data is uploaded and parsed correctly
  • ✅ Model selection is configured
  • ✅ All required fields are present

    • Running the Forecast

    1. Review the specification summary

    2. Click "Run Forecast" button

    3. Progress indicator shows forecast execution

    4. Results page appears when complete

    What happens during execution:

    1. Data profiling (statistical analysis)

    2. Model selection (if Auto mode)

    3. Model training

    4. Forecast generation

    5. Quality diagnostics

    6. Results compilation

    Typical execution time:

  • Small datasets (< 10 series): 5-15 seconds
  • Medium datasets (10-100 series): 15-60 seconds
  • Large datasets (100+ series): 1-5 minutes

    • Buttons

    Button Action
    **Previous** Go back to Step 4
    **Run Forecast** Execute the forecast
    **Save Specification** Save without running (for later)
    **Cancel** Close wizard without saving

    Tips

  • **Review carefully** - Check all settings before running
  • **Save specification** - Save your configuration for future use
  • **Be patient** - Larger datasets take longer to process
  • **Check data quality** - Ensure data preview looks correct

    • Common Issues

    Issue Cause Solution
    "Validation failed" Missing required fields Go back and complete all steps
    Forecast takes too long Large dataset Normal for large datasets, wait for completion
    "Insufficient data" Not enough historical data Provide at least 2× seasonal period of data

    Results View

    [Screenshot: Results Page]

    After running a forecast, you'll see:

    Forecast Summary

    Field Description
    **Run ID** Unique identifier for this forecast run
    **Status** COMPLETED, RUNNING, or FAILED
    **Business Target** What was forecasted
    **Series Count** Number of time series forecasted
    **Models Used** Which models were applied
    **Execution Time** How long the forecast took

    Forecast Chart

    Visualization showing:

  • **Historical Data** (blue line) - Your actual past values
  • **Forecast** (green line) - Predicted future values
  • **Confidence Intervals** (shaded area) - Upper and lower bounds

    • Chart Controls:

  • **Aggregation**: View raw, daily, weekly, or monthly aggregates
  • **Zoom**: Click and drag to zoom into specific periods
  • **Export**: Download chart as image

    • Profiling Summary

    Statistical analysis of your data:

    Metric Description
    **Mean** Average value
    **Trend** Upward, downward, or stable
    **Seasonality** Detected seasonal patterns
    **Stationarity** Whether data is stationary
    **Data Quality** Overall data quality assessment

    Diagnostics

    Quality assessment and recommendations:

    Aspect Description
    **Overall Quality** Good, Fair, or Poor
    **Recommendations** Suggestions for improvement
    **Model Performance** How well models fit the data

    Export Options

    Format Button Use For
    **Specification** "Download Spec" Re-running forecasts later
    **Results** "Download Results" Sharing forecasts with stakeholders
    **Diagnostics** "Download Diagnostics" Quality analysis
    **Profiling** "Download Profiling" Statistical details

    Progress Indicator

    The step indicator at the top shows your progress:

    ```

    [Step 1] ─ [Step 2] ─ [Step 3] ─ [Step 4] ─ [Step 5]

    ● ● ● ● ●

    completed completed completed completed active

    ```
    State Appearance
    Completed Filled circle, checkmark
    Active Highlighted, current step
    Upcoming Empty circle

    Saving Your Work

    Your work is saved when you:

  • Click **"Save Specification"** in Step 5
  • Complete a step and click **"Next"**
  • The wizard auto-saves periodically

    • Going Back

    You can always go back to previous steps:

  • Click **"Previous"** button
  • Click on a completed step in the progress indicator

    • Keyboard Shortcuts

    Shortcut Action
    `Enter` Confirm / Next
    `Escape` Cancel / Close
    `Ctrl+S` Save

    Troubleshooting

    Wizard Won't Open

    Cause Solution
    Browser issue Refresh page
    JavaScript disabled Enable JavaScript
    File path issue Check file path is correct

    Data Upload Fails

    Cause Solution
    Wrong file format Use CSV or JSON
    Invalid date format Convert dates to YYYY-MM-DD
    Missing required fields Ensure unique_id, ds, and y are present
    File too large Split into smaller files or optimize data

    Forecast Fails

    Cause Solution
    Insufficient data Provide at least 2× seasonal period
    Data quality issues Check for missing values or outliers
    API offline Verify services are running on ports 8020 and 8021
    Invalid specification Review all steps and validate

    Poor Forecast Quality

    Cause Solution
    Not enough historical data Collect more data points
    Data quality issues Review diagnostics and clean data
    Wrong frequency Match frequency to data spacing
    Model mismatch Try different models or use Auto mode

    Quick Reference

    Step Summary

    Step Required Input Minimum
    1 Business target Select one target
    2 Frequency, horizon, confidence All three fields
    3 Data file Valid CSV or JSON
    4 Model selection mode Auto or Manual
    5 Validation Must pass validation

    Button Reference

    Button Where Action
    Cancel All steps Close without saving
    Previous Steps 2-5 Go back one step
    Next Steps 1-4 Save and proceed
    Run Forecast Step 5 Execute forecast
    Save Specification Step 5 Save without running

    Next Steps

  • **[Quick Start Guide](QUICK_START.md)** — Step-by-step tutorials
  • **[Concepts Guide](CONCEPT_GUIDE.md)** — Understanding forecasting concepts
  • **[User Guide](USER_GUIDE.md)** — Complete reference

    • Support

  • Click the **?** icon in any step for context help
  • Email: support@planningpowertools.com