When and how much are developers working?
Inspired by this year’s Stack Overflow Developer Survey results we’ve investigated Samebug’s site usage data. We do not have a question similar to the one in the SO survey about when developers wake up or how much time they spend daily on their computer but we do know the local time of our visitors who come to find solutions to their bugs.
The SO study came to the conclusion that the majority of the developers wakes up between 6 AM and 8 AM. Based on local visit times we can also conclude that developers generally start working at 8 AM, they have a peak at 11, then after a dinner (typically between 12 and 14), comes another peak at 15 PM. They generally stop working at 5 PM on but some of them keep going until midnight.
We have also made a comparison of timetables in different countries. The site visits show a country distribution quite different from that of the developer survey. Thus we have chosen the first 20 countries with most visitors in our database and compared only those countries to the survey’s data. We have calculated an average wake-up time from the WakeTime variable and an average visit-time from our LocalTime variable and ranked the countries by both. The comparison is illustrated in a scatter plot. The Spearman correlation coefficient, a measure for strength and direction of a correlation between ranked variables, is 0.48.
This moderately strong relation can be seen on the diagram as well. Most countries are in the bottom left and the upper right side. For example, Swiss developers wake up early according to the developer survey and they also turn up soon on Samebug every day. In contrast, developers from Ukraine tend to start later in the day. Russian developers seem to behave differently or they are just simply fond of Samebug when delving into debugging.
We were also interested in what these data sources tell us about how much and how intensively developers work. We calculated some site-usage intensity based on the behavior of our returning visitors and tried to point out some connection with the developer survey’s questions: “how much time developers spend on a computer?” and “how often do they skip meals to be productive?” (which, by the way, don’t seem correlated with each other). We looked at the country averages to help us join the data sources and made the rankings like before for both data sets. We took into account only the 20 countries with the most visits.
We found that in those countries where the developers use Samebug intensively, they also tend to skip their meals quite frequently (Spearman coefficient 0.58). Chinese developers won silver medal in one and bronze medal in the other ranking. Albeit we are not sure if it is just another spurious correlation or it means that debugging with Samebug is worth skipping a meal.
