Resources to Download

• Postman Collection
• Jupyter Notebook
• Excel File
• Presentation Slides

Transcript: Getting Started with the ForecastWatch API

Introduction (0:00-3:40) In this session, we explore the ForecastWatch REST API. This API powers the entire user interface and allows users to programmatically pull data, generate reports, and conduct deep analysis using tools like Postman, Excel Power Query, and Python/Jupyter Notebooks.

API Authentication & Setup (6:04-8:50) To get started, log in to your ForecastWatch account, navigate to Tools > API Portal, and generate an API key.

• Best Practice: Create a unique key for each application (e.g., one for Excel, one for your Python scripts). This enhances security—if one key is compromised, you can delete it without affecting your other workflows.

Working with Postman (9:16-19:33)

• Base URL: api.forwatch.com
• Header: Use Authorization: Bearer.
• Workflow: Chain requests by fetching reference data first (e.g., v1/reference/months), then use the returned IDs to query specific endpoint data like v1/insights/monthly.

Excel Power Query (19:34-24:06) Excel users can pull live API data directly using Power Query. By navigating to Data > Get Data > From Other Sources > Blank Query, you can use the advanced editor to input your bearer token and API calls. This enables refreshable, dynamic data sets for your spreadsheets.

Advanced Analysis: Python & Jupyter Notebooks (24:39-44:07) Using Google Colab or local Anaconda environments, you can perform sophisticated analysis.

• Use Case: The demo showcases how to identify “chronic offenders”—weather stations where a model is systematically miscalibrated (e.g., consistently running 10 degrees cold).
• Methodology: By pulling the “Bold and Wrong” quadrant data over a 14-day period and using Pandas to group by station, you can isolate specific locations that require attention.

Best Practices Summary (45:32-49:28)

1. Chain Requests: Always fetch IDs from reference endpoints before querying data endpoints.
2. Security: Never hardcode API keys. Use environment variables (Postman) or Secrets (Colab).
3. Filtering: Use API parameters to filter data at the source to keep requests lightweight.

For further assistance or documentation, contact support@forecastwatch.com or visit the live endpoint catalog at app3.forecastwatch.com/platform/api.