What is this?
This script is used to "measure the weather" on Hacker news.
Just like in real life we don’t simply look at the last temperature at 11:59 and say, "This is how the day was." That would be wrong and inaccurate. We take multiple readings on multiple variables per day to get a more complete story.
That is the idea with this little project. Here, we will mesaure the weather on Hacker News across a number of variables. My hope is that over time the collection and publication os this data would prove valuable, or at least interesting to some folks. The variable we capture are:
- number of comments
- "karma score"
- ranking (the link’s position in the list)
This is a python script and docker image which will collect bookmark rankings from Hacker News ( http://news.ycombinator.com ).
We want to be polite, so we only pull data once every two minutes for the first five pages.
- Python 3+
- Docker / Docker compose (optional)
NOTE: you do not have to have docker to get this thing going.
NOTE: I do not believe that postgres is required — it is just what I chose.
- Setup your database
- Build your database url string and put that into you
.envfile eg – postgres://user:pass@host/dbname
python db.pyto setup the tables in your DB
docker-compose up -dto run in backgroundo
- postgresql or other db (mysql, mysq lite) as configure in docker-compose.yml
- some visualtion would be nice. Right now the data just live in the DB. Maybe a flask app?
- Perhaps it might be interesting to do this with reddit or another social news site
- Public a live API feed / data dumps of the HN feed for anyone to analyze
转载本站任何文章请注明：转载至神刀安全网，谢谢神刀安全网 » Hn_weather, generating time-series “weather” data of Hacker News