I built a tool that tells you what developers are actually learning — not just Twitter hype

Source: Dev.to

The Problem

Tech hype is asymmetric. A single viral tweet can make a niche tool look dominant, while slow‑burning technologies that millions of developers are quietly adopting get ignored.

The only way to cut through this is to look at behavior, not opinion:

• What are developers actually building? (GitHub)
• What are they writing about? (Dev.to)
• What are they discussing? (Hacker News)

What trend‑radar does

pip install trend-radar
trend-radar

It fetches trending data from all three sources in parallel, normalizes the signals, weights them (GitHub 50 %, Dev.to 30 %, HN 20 %), and outputs a ranked table:

 Technology    GitHub  Dev.to  HN    Score
 Rust          ████    ███     ███   89
 Bun           ████    ████    ██    84
 Deno          ███     ███     ██    71

You can also get raw JSON for scripting:

trend-radar --json --top 10

How it works

Three independent collectors run concurrently via ThreadPoolExecutor:

• GitHub collector – scrapes github.com/trending and counts repos per programming language/tool using BeautifulSoup.
• Dev.to collector – hits the Dev.to public API (/api/tags) and pulls tag post counts as a proxy for what developers are writing about.
• Hacker News collector – uses the HN Firebase API to fetch top‑story titles and counts keyword mentions across them.

The analyzer normalizes each source to a 0–100 scale, applies the weights, and returns a sorted TechScore list. The display layer renders it as a rich terminal table.

What I learned

• Scraping is fragile, APIs are gold. The GitHub scraper works today but GitHub changes its HTML occasionally. The Dev.to and HN APIs are stable and versioned—much more reliable.
• Signal weighting matters more than collection. Getting the data is the easy part. Deciding how much to trust each source (and why) is the actual design problem.
• ThreadPoolExecutor makes I/O‑bound fetching trivial. Three HTTP calls that each take 2–3 seconds run in under 4 seconds total. The core code is only about 10 lines.

Try it

pip install trend-radar
trend-radar --top 20

Source: github.com/LakshmiSravyaVedantham/trend-radar

Would love to hear what you think—and what sources you’d add to make the signal stronger.