I have a weekly dataset. I use this code to plot the causality between variables. Stata shows the number of weeks of each year on the X-axis. Is it possible to show only year or year-month instead of year-week on the X-axis?
generate Date =wofd(D)
format Date %tw
tsset Date
tvgc Momentum supply, p(3) d(3) trend window(25) prefix(_) graph
The fact that you have weekly data is only a distraction here.
You should only use Stata's weekly date functions if your weeks satisfy Stata's rules:
Week 1 starts on 1 January, always.
Later weeks start 7 days later in turn, except that week 52 is always 8 or 9 days long.
Hence there is no week 53.
These are documented rules, and they do not match your data. You are lucky that you have no 53 week years in your data; otherwise you would get some bizarre results.
See much detailed discussion at references turned up by
search week, sj.The good news is that you need just to build on what you have and put labels and ticks on your x axis. It's a little bit of work but no more than use of standard and documented label and tick options. The main ideas are blindingly obvious once spelled out:
Labels Put informative labels in the middle of time intervals. Suppress the associated ticks. You can suppress a tick by setting its length to zero or its colour to invisible.
Ticks Put ticks as the ends (equivalently beginnings) of time intervals. Lengthen ticks as needed.
Grid lines Lines demarcating years could be worth adding. None are shown here, but the syntax is just an extension of that given.
Axis titles If the time (usually x) axis is adequately explained, that axis title is redundant and even dopey if it is some arbitrary variable name.
See especially https://www.stata-journal.com/article.html?article=gr0030 and https://www.stata-journal.com/article.html?article=gr0079
With your data, showing years is sensible but showing months too is likely to produce crowded detail that is hard to read and not much use. I compromised on quarters.