Why some social media posts fall into a pit of oblivion while others shine brightly is in great part a question of timing. A new study from Klout finds that, unsurprisingly, posts that land sometime during the workday tend to receive the most reactions — and that beyond that, it’s a matter of geography.

The study, by Nemanja Spasojevic, Zhisheng Li, Adithya Rao, and Prantik Bhattacharyya, sampled 144 million Twitter and Facebook posts and around 1.1 billion reactions to those posts — retweets, comments, likes, and so forth. In the U.S., people on social media in New York and San Francisco tend to react more in the first half of the work day. In Paris, reactions peak in the later half of the work day; in London, near the end of the work day. The most notable exception: Tokyo, where reactions on social media peak twice in the day, both outside of working hours.

As Klout’s blog post on the data put it:

User schedules need to be personalized for maximum engagement, and using a generalized schedule based on regional averages is limited in effectiveness. Why? Because any user’s audience is typically spread across various locations. So, when you tweet from Dallas, Texas it reaches your audience that is in the same city/time zone as you, but it doesn’t arrive at the prime time for reactions for any of your followers who are in different locations around the country or globe. Thus, the likelihood of them reacting is much lower.

On weekends, Twitter activity drops significantly across the board, though Facebook is still used fairly regularly (and most consistently on Sundays), the study found. Timing, though, is everything:

We find that a majority of reactions occur within the first two hours of the original posting time on most networks. Audience behavior differs significantly on different networks though, with Twitter having larger reaction volumes in shorter time windows immediately after the post (50% of reactions within the first 30 minutes), and as compared to Facebook which reaches 50% of reactions after two hours.

The researchers have made their dataset available publicly if you want to check their work — or just want an anonymized set of a billion or so social media posts to play with over the weekend.