Sure thing. Pick the lightest sensor board you can find—DHT22 or BME280, both cheap and low power. Hook it up to the Pi’s GPIO, read raw bytes, then use a small, pre‑compiled C library to parse the data instead of a Python wrapper. That cuts the loop time by almost a factor of two. If you’re logging, write directly to a compressed CSV in a single syslog entry; no extra Python overhead. Once you’ve got a clean stream, you can compute moving averages with a fixed‑size ring buffer in C, then expose a minimal REST endpoint in Go for the UI. That’s the most efficient pipeline I’ve seen for a Pi‑based weather station.

Sounds good. The BME280 will give you cleaner numbers and a quicker cycle, but keep an eye on the overshoot it can do at the edges. For the tiny HTTP service, one goroutine with a short keep‑alive timeout will do, no need for a full‑blown framework. Just remember to add a watchdog on the Pi; a stuck sensor can leave the whole thing in a loop waiting for data that never arrives.