Peak time for being on the internet in Britain is 21:00 on Wednesdays, the Office for National Statistics (ONS) suggests.
By contrast 04:45 on weekdays is the quietest time, it says.
The ONS has started analysing internet traffic volume for the first time, using publicly available data from one of Britain's largest internet exchanges.
However, it acknowledges the early results are not conclusive.
One expert told the BBC the data used had a significant limitation because it was unlikely to have included box-set streaming.
The study made use of data from the London Internet Exchange (Linx), which connects internet service providers and content delivery networks.
It operates exchanges in London, Edinburgh, Cardiff and Manchester, but some of the data was extrapolated from the London area alone.
Andrew Ferguson, from news website, ThinkBroadband, told the BBC that while the study had merit, often online box-set viewing will not be included in Linx data as this is delivered in a different way.
"This is not the full story but when combined with other data can reveal trends, rather than absolute figures," he said.
While the ONS said it was still working on the statistics, it found that internet traffic in the UK:
- falls between 16:00 and 18:00 with its lowest point at 18:00 on weekdays (not weekends)
- increases in winter, decreases in summer
- is quietest between 02:00 and 07:00
- peaks at 21:00 on weekdays, particularly midweek
- is generally quieter at weekends and bank holidays
- peaks at 13:00 during weekends
- Saturday night is lowest data flow
The ONS made comparisons with other data such as road traffic information - for example, matching an increase in traffic on the M25 motorway around London during the time that internet data traffic was lower, as people commuted home after work.
This approach also revealed some interesting behaviours around special events: there was not a significant spike in data flow during online shopping sale cyber- Monday, and during the England v Croatia football World Cup semi-final in 2018 there was more data exchanged during half time and before the extra time at the end of the match than there was during game itself.
"Insights on large-scale events such as adverse weather and sports could potentially help to provide measurements of economic impact over time," wrote Philip Stubbings, lead data scientist at the ONS Data Science Campus.