In the midst of the pandemic (2020), i worked for a non-profit....and they published all manner of important economic research data on different strata of incomes, families who struggle putting food on the table, etc. All data that they sell to governments for policy work (its that great, valuable, and expertly published) and they also sell it to other research firms, NGOs, etc. (Side note: when i worked for them, they were severely under-pricing their data for way too low prices for how others used, referenced the data) Anyway, among the steps involved in reviewing the reports, I automated link checking of urls used within their reports. The report content lived in PDFs, so i automated grabbing all content from PDFs (not fun!), then picking out urls from the massive copy, then running link checker scans, finally producing a report of any links that were broken. The overall authoring of the reports encompassed many, many months of work - by top-notch research teams - and the design and production of said reports also (sadly) took months to perform. Well, when i automated even just the link checking portion, i reduced their work from many weeks (sometimes 1 or 2 months) down to 1 or 2 days. I left before i could extend my automation, so it still required a human to trigger some steps. But, if i still worked there, i planned to automate far more, and start moving upstream for other areas (like scaling out their report designs, etc.).
Was it some hacker-level thing or rocket science? No way. But, all in all, it was a neat use of python, python pdf modules, some minor bash/cli scripting, etc. I learned a bunch of fun stuff, and the non-profit saved tons of time. I still feel so good when i reflect on that little project!