Please note that a similar question was asked on 13 February:
<http://www.hsph.harvard.edu/cgi-bin/lwgate/STATALIST/archives/statalist
.0902/date/article-543.html>
And answered on the same day by Ulrich Kohler and myself:
<http://www.hsph.harvard.edu/cgi-bin/lwgate/STATALIST/archives/statalist
.0902/date/article-553.html>
<http://www.hsph.harvard.edu/cgi-bin/lwgate/STATALIST/archives/statalist
.0902/date/article-557.html>
Despite what you say, it is not obvious that the graph form you seek is
likely to be especially efficient or effective in showing changes over
time.
You can get sum variables
foreach v in variable1 variable2 variable3 variable4 {
egen sum_`v' = sum(`v'), by(timeperiod)
}
And then you can plot the sums directly by
line sum_* timeperiod
Nick
[email protected]
David Airey
Mess around and you'll get it, for example, see:
clear
input time var_4 var_5 var_6
1 10 20 30
2 20 30 40
3 30 40 50
end
graph bar var_4-var_6, over(time)
reshape long var_, i(time) j(variable)
graph bar var_, over(time) over(variable)
clear
input time response var
1 10 4
2 20 4
3 30 4
1 20 5
2 30 5
3 40 5
1 30 6
2 40 6
3 50 6
end
graph bar response, over(time) over(var)
On Mar 25, 2009, at 8:30 AM, Kate McAllister wrote:
> Does this mean I have to rearrange my data set? At the moment column 1
> contains my 'timeperiod' variable and then the other columns are my
> other
> variables, with a total of three lines in my table. Is the answer to
> have
> the timeperiods as different columns and the variables as different
> rows?
> If so, is there a nice simple way of doing this (I have quite a few
> variables)
David Airey
>> What you wanted is to have your times (ordered correctly) over your
>> variables, not your variables grouped by time, which is what you got.
>> The help file for graph bar talks about what you can expect under
>> different requests. You have to scroll way down to ``remarks'',
>> specifically where it gives pictures of year and division.
>> On Mar 25, 2009, at 6:58 AM, Kate McAllister wrote:
>>> I want to create a bar graph displaying several variables over 3
>>> different
>>> time periods. If I type:
>>> graph hbar (sum) variable1 variable2 variable3 variable4,
>>> over(timeperiod)
>>>
>>> then I get a graph that displays all the variable values for
>>> timeperiod1
>>> clustered together, then all the variable values for timeperiod2,
>>> and then
>>> for timeperiod3.
>>>
>>> What I actually want is a graph the displays the variable1 values
>>> over the
>>> three time periods, THEN the variable2 values over the three time
>>> periods,
>>> then the variable3 values... and so on, all on one graph.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
