2026 Data Import Playbook: Loading Economic Data with Power Automate

Introduction

Most people who track economic indicators do not want to run a data pipeline. They want yesterday’s unemployment figure or the latest CPI release to show up in a spreadsheet, a dashboard, or a Teams channel without babysitting a server. In 2026, the most common way non-engineers solve this is Microsoft Power Automate, because it ships inside the Microsoft 365 licenses many organizations already pay for and it can call a public API on a schedule without any hosting of your own.

This guide walks through a repeatable pattern for importing economic data with Power Automate. It is deliberately simple: a scheduled flow, an HTTP request to a public data source, a small parse step, and a write to a destination you already use. The goal is a load process you can set up once and walk away from.

💡 Tip: Power Automate is a good fit when your data updates daily or less often and your volumes are small (a few hundred rows per run). For high-frequency or large imports, a scheduled job in GitHub Actions or a Cloudflare Worker will be cheaper and more predictable. We cover that trade-off in a companion guide.

The core pattern

Every import flow in this playbook has the same four stages, regardless of the source:

1. Trigger — a Recurrence trigger that fires on a schedule (for example, every weekday at 8:30 AM in your timezone).
2. Fetch — an HTTP action that calls the data source’s API and returns JSON.
3. Shape — a Parse JSON or Select action that pulls out only the fields you care about.
4. Land — an action that writes the result somewhere useful: a row in Excel, a SharePoint list, a Dataverse table, or a message in Teams.

Keeping every flow to these four stages makes them easy to copy. Once you have one working source, adding a second is mostly changing the URL and the fields.

Example: importing a FRED series

The Federal Reserve Economic Data (FRED) API is a clean source to start with because it returns simple JSON and offers hundreds of thousands of series. You will need a free FRED API key, which you request from the St. Louis Fed’s developer portal.

Set up the flow as follows. Add a Recurrence trigger set to run once a day. Add an HTTP action with the method set to GET and a URL in this shape:

https://api.stlouisfed.org/fred/series/observations?series_id=UNRATE&api_key=YOUR_KEY&file_type=json&sort_order=desc&limit=1

That request asks FRED for the single most recent observation of the unemployment rate series (UNRATE). Add a Parse JSON action and point it at the body of the HTTP response so Power Automate understands the structure. Finally, add an Add a row into a table action (Excel Online) that writes the observation date and value into a tracking workbook.

The result is a workbook that gains one fresh row every morning with no manual effort.

📌 Note: Never paste your API key directly into the URL field if you can avoid it. Store it as an environment variable in a Power Platform environment, or in Azure Key Vault, and reference it from the flow. A key sitting in plain text inside a flow definition is the most common avoidable mistake we see.

Handling more than one series

Tracking a single indicator is rarely the goal. To import several series, you have two clean options. The first is to duplicate the HTTP and write steps for each series — fine for three or four indicators, tedious beyond that. The second, and the one we recommend once you pass a handful, is to drive the flow from a list.

Create a small table of the series you care about (series ID, friendly name, destination). Use an Apply to each loop over that table, and inside the loop run the same Fetch and Land steps using the current row’s series ID. Adding a new indicator then becomes a one-row edit rather than a flow rebuild. This is the same idea as a configuration file in code, expressed in a no-code tool.

Choosing a landing spot

Where you write the data matters more than how you fetch it, because the landing spot is what your readers and dashboards actually touch.

Excel Online is the lowest-friction choice and works well for personal tracking or small teams. It struggles once a workbook grows past tens of thousands of rows. SharePoint lists handle larger volumes and give you views and filtering for free, at the cost of clumsier formulas. Dataverse is the most robust option and integrates directly with Power BI, but it requires a Power Apps or premium Power Automate license, so check your plan before committing.

For most people getting started, Excel Online is the right answer. You can migrate the landing step to SharePoint or Dataverse later without touching the fetch logic.

What this approach is good and bad at

Power Automate earns its place because it removes the two things that make people abandon data tracking: hosting and maintenance. There is no server to patch, no deploy step, and the flow keeps running whether or not you are paying attention. For daily economic indicators, that reliability is worth more than raw efficiency.

Its limits are real, though. Power Automate’s HTTP connector is a premium action on many license tiers, the per-flow run limits can bite if you loop over many series, and debugging a failed run inside the web designer is slower than reading a log file. If you find yourself fighting those limits, it is a signal that your import has outgrown the tool — which is a good problem, and one the next guides in this series address.

Conclusion

A four-stage flow — trigger, fetch, shape, land — is enough to keep a meaningful set of economic indicators flowing into a spreadsheet or list that your team already uses, with no infrastructure to own. Start with one FRED series writing into Excel Online, confirm it runs unattended for a few days, then expand by driving the flow from a configuration table. The aim throughout is a load process that survives your inattention, which is exactly what a long-running tracker needs.