Fellow news geek Greg! (ツ)
If you haven't looked into it, something you might either love or hate are Google Alerts (introductory info from Google here). Any number of alerts can be individually created, configured, and managed using Google search strings and parameters for frequency of alerts, types of sources used, region of interest, etc. While the help info and creation/management screens only talk about sending email alerts, when you receive your first one you'll find, at the bottom of the message, a link titled "Receive this alert as RSS feed." Yep; you're provided a unique feed address for your specific alert. Voila!
I utilize some of that in a marginally effective news aggregation service I publish in conjunction with one of my one-name studies. It has the very catchy domain name TheTribune.news, subtitled "Your Daily Genealogy, Genetics, and One-Name Study News."
I use the Paper.li aggregation service for that, and I say it's "marginally" effective because it almost always needs some tweaking after each new edition is pushed to the Web at around 9:30 Eastern each morning. I can move, delete or add content items after it comes out. There is an option to do full curation--meaning compiling each "edition" in a draft mode before publishing--but that means every edition has to be pushed manually...resulting in much more work and variable publication times/dates.
About 10% of the content ends up being specific to the Threlkeld One-Name Study (includes pulling in a Google Alert for mentions of the surname, which also grabs mentions of the town of the same name in northwest England's Lake District), but I also pull in, as RSS sources, many of the well-known genealogy bloggers like Blaine Bettinger and Debbie Kennett and Dick Eastman, as well as a Twitter genealogy list I created that includes multiple sources (e.g., family history associations, individual genealogists, organizations and companies like ISOGG, Ancestry, FamilySearch, the National Archives, NEHGS).
Paper.li includes its own article-search utility that's pretty broad: it seems heuristics-based and, while you can set specific search word/term exclusions and blacklist sites, it really does reach far and wide, sometimes coming up with unexpected results. I'd say that most of the five or so minutes I spend almost every day editing TheTribune.news aggregation is spent on things pulled in by the Paper.li engine; the remainder would be from very local news from the BBC about Cumbria, England (outside of a very few, a local ambulance vehicle being out of service isn't all that interesting). The "paper" can include up to 25 different source entries, but using RSS as well as account/list targets for Twitter, Facebook, Google+, and YouTube, plus the Paper.li proprietary search tool, it provides a pretty eclectic breadth.
It's kind of a fun exercise and, if nothing else, it gives me my own, daily newspaper to glance through. ;-)
