-
Notifications
You must be signed in to change notification settings - Fork 7
Home
Welcome to the tt-history wiki!
Purpose of tt-history is to help people to understand that how much time trending topics of Twitter are talked. Although Twitter shows us current trending topics, with this project we can see all trending topics within last 24 hours or specific date. Further we can compare trending topics in terms of how much time they are talked. At this point we can raise a question that is there any relationship between the trending topics and can visualization help us to understand what's going on there?
Project runs on Google App Engine and for visualization I used d3.js. I chose App Engine because of it has free plan. However currently I switched to paid plan and I pay how much resource I use.
Every 10 minutes tt-history fetches trending topics and stores response to the data-store with request date and time. Therefore you can learn trending topics at specific date (after 2013) (with hour and minute resolution). Because of this methodology you have to store lots of things on database. For every 10 minutes it inserts 10 entry for specific region (or woeid). For one day you have 1440 entry. When someone enters tt-history 1440 entries are fetched and summed. But there are caching mechanism, therefore actually not every time that amount of entries are fetched. Showing last week's and lost month's trending topics are hard because of this method. By the way for tt-history stores only trending topics of Turkey and World, and database size has reached to 1.7GB at February 28.