How Remote Work Has Grown – and Shrunk – Since Covid

lumpa1 pts0 comments

How Remote Work Has Grown — and Shrunk — Since Covid – Ari Lamstein

Skip to content

How Remote Work Has Grown — and Shrunk — Since Covid

Ari Lamstein

May 18, 2026

Remote work surged during Covid — and while it has declined since, it’s still far above pre‑pandemic levels. I just updated my Covid Demographics Explorer with the latest ACS data, and the national trend is striking:

Remote work more than tripled between 2019 and 2021, rising to nearly 28 million people at the height of the pandemic. Since then it has edged down each year, but only modestly. Even today, at about 22 million, it remains roughly 2.5 times the pre‑Covid level.

The app now lets you generate this same graph for every state, as well as for counties and cities with populations of at least 65,000. See how the trend looks where you live.

Exploring Local Trends

I also added a "Compare Years" tab that lets you see which locations saw the biggest change in remote work between any two years. The national trend tells one story, but the local data tells another: the rise and fall of remote work played out very unevenly across the country. Below I run this analysis twice: first for the national increase from 2019-2021, and then for the gradual decline between 2021 and 2024.

The Remote Work Spike: 2019-2021

Between 2019 and 2021, the location that increased the number of remote workers the most was Sunnyvale, California. The number of remote workers there increased almost 11x in two years, from an estimated 3,235 to 38,319. Sunnyvale is in the heart of Silicon Valley, and tech companies were among the fastest to adopt remote work, which helps explain this result:

The scatterplot also shows the broader pattern: most locations cluster between a 150% and 300% increase in remote work during this period. That makes Sunnyvale’s nearly 1,100% jump stand out even more — it’s an order of magnitude beyond the national norm.

Interestingly, only one location in the entire dataset saw a decrease in remote work during this period: Rice County, Minnesota (-7.5%). It’s the lone point below zero on the chart, and I don’t have a clear explanation for it.

The Remote Work Decline: 2021-2024

When we run this same analysis for 2021–2024, we see a very different result: Sunnyvale’s remote workforce shrank by 67.2%, the largest drop in the dataset. This means that Sunnyvale saw both the largest increase between 2019 and 2021 and the largest decrease between 2021 and 2024:

The scatterplot also shows how different the overall pattern is in this period. Instead of large increases, most locations cluster between a 10% and 30% decline in remote work — a sharp contrast with the 2019–2021 graph, where nearly every location saw a substantial increase.

Against this backdrop, Sunnyvale’s 67% drop stands out as an outlier. The likely explanation is the wave of return‑to‑office mandates that swept through the tech industry during this period. The two other largest decreases also happened in Silicon Valley: the city of Fremont (–61%) and Santa Clara County (–56%).

At the other end of the distribution, the few places that saw increases tend to be warm‑weather, high‑amenity destinations: Marion County, Florida (69%), Collier County, Florida (65%), and Maui County, Hawaii (57%) saw the largest gains. These increases may reflect people with remote‑work jobs relocating to places with natural beauty and a high quality of life — a very different dynamic from the employer‑driven declines we see in Silicon Valley.

Conclusion

Three years after the peak, roughly 22 million Americans still work from home — more than double the pre-pandemic baseline. But the story is more complex than a single national number: a dramatic surge, an uneven retreat, and striking differences across the country. How does your corner of the country fit in?

The new version of the Covid Demographics Explorer makes it easy to explore these patterns yourself. In addition to remote‑work trends, you can examine changes in population, median household income, median rent, and public assistance. Analyze your own location.

This app was built in Python with the Streamlit framework. I teach Streamlit for O’Reilly — and if you’d like to learn to build apps like this yourself, I offer a free 7-day email course. Sign up in the form below.

Share this:

Share on Facebook (Opens in new window)<br>Facebook

Share on X (Opens in new window)

Share on LinkedIn (Opens in new window)<br>LinkedIn

PrevPreviousA Web App for Exploring Foreign‑Born Population Trends

Ari Lamstein

I’m a software engineer who focuses on data projects.<br>I most recently worked as a Staff Data Science Engineer at a marketing analytics consultancy. While there I developed internal tools for our data scientists, ran workshops on data science and mentored data scientists on software engineering.<br>I have also created several open source projects.

Latest Blogs

May 18, 2026

How Remote Work Has Grown — and Shrunk — Since Covid

April 27, 2026

A Web App...

remote work covid between data since

Related Articles