nloui

Looking forward to tracking this project. I totally get it.

Very cool. Also just signed up for your product beta.

I may open source it. It's pretty easy to get started if you leverage pre-existing libraries for Markov chains and Twitter ingestion.

Jane?! That is a small world and a great memory! I used to be a huge ColdFusion fan.

Sure, definitely. Or just retrain Parrot on VC Medium articles and have it write itself :-)

Born and raised!

Sure, I should be able to throw it onto a repo later.

They generate on a cron job every 15 minutes and I also pass the site through a Cloudflare cache because I'm cheap and the actual code runs on a $5 DigitalOcean box :-)

Hmmm I hadn't thought of that. That's a great point.

Yep!

Hah oh jeez.

Ugh my relationship with Spectrum

Hah that's pretty good

Perfect - starting the AngelList syndicate right now. Haha

It's a markov chain trained on around 65,000 tweets from around 50 VC's.

Haha I'm sure it's somewhere in the middle. Never said it was the smartest algorithm. If you hover over the VC's name, it'll show the actual tweet that was sampled.

Parrot.vc is a bot trained on tens of thousands of VC tweets that uses predictive text to generate amazing, new startup advice. Share your favorites!

Gavin Belson - hit me up, this is the perfect acquisition for Hooli.

PS - you can follow Parrot.vc on Twitter as well (https://twitter.com/parrot_vc)

We're not quite real-time but this is in our wheelhouse https://www.civicfeed.com/

Hi HN,

Explore News is a side project I built that helps you visualizes how people and organizations are connected through a graph.

As part of the work we do at CivicFeed, we've built a pipeline of news/social/government datasets. From there, we run our own entity extraction and entity disambiguation algorithms.

Over the weekend, I was curious about how we could use this data to show how news stories connect people and organizations.

So to play with it, type in a query (or click one of the suggestions). It'll return the top entities most associated with that query. Then click another gray dot. When you do, it'll return the top entities associated with that query then connect any dots.

Keep exploring from there.

Yup, totally. Was just curious if someone had made this process just a little bit easier.

If you're looking for more news sources, happy to hook you up with some extra API credit (https://developers.civicfeed.com/)

This is great. Setting up proper staging environments has been on our procrastination list for a long time and as a small team, we haven't invested engineering hours for a real setup yet. If this simplifies the process so much that we barely need to think about it, I totally get the value.

Thanks for the comment! Because of the vast amount of data, we saw an opportunity to use natural language processing to better parse and understand the millions of documents that existed. This allows for faster discovery of information, more relevant data being served, and a better understanding of the text.

Hey guys! We were super frustrated not being able to understand the laws, organizations, and people that affected us. With over 8,500 bills/resolutions in front of Congress and only 7% that pass on average - there was a lot to unpack (not even including the states!). So, we used our background in AI/NLP to build a product that brings government affairs into the modern world.

cool! have you (or have you considered) open sourcing this?

I just rejoined Thislife.com, now part of Shutterfly. For Smugmug lovers, I also created Smugsync which will sync iPhoto to Smugmug => smugsync.net