[Nico Stappen]

so the command is

graph twoway line tradevalueus period_new if tradeflow=="exports" & partner=="belgium", c(L)

Than the graph is:

Hope I'm getting closer to my solution step by step, but this graph doesn't seems to be the right.
Thanks for your prompt replies helps me not to be completely frustrated directly

[Nick Cox]

Note however that your period_new axis labels make no sense.

It's possible to reconstruct what you did.

You wrote something like

format %tm period_new 

on the grounds that period_new is a monthly date variable. But Stata's idea of a monthly date variable is strict. It is that 0 indicates January 1960. In your case 0 implies some other date.

For most purposes mdate would be better for graphing the data than period_new. If you want period_new shown as such, with values 1 up, then take off the monthly date format by assigning

format %4.0f period_new

[Nico Stappen]

.

[Nick Cox]

I can't see the graph in #18. Only a .png attachment will work.

The option syntax looks right in #18, but #17 implies that period_new is not the best choice here.

[Nico Stappen]

.

[Nico Stappen]

*Data import*


import excel "/Users/nicolasstappen/Desktop/Nico/Studium/Master/5. Semester/Masterarbeit/Daten/all Data in one file nur ab 2019.xlsx", sheet("Trade values") firstrow case(lower)

codebook // entweder nur codebook oder codebook varname

drop classification // no information and/or variation
drop perioddesc // no info that is not within Period already
drop reportercode // no information and/or variation
drop partnercode // no added information
drop tradeflowcode // no added information

replace tradeflow="exports" if tradeflow=="Exports"
replace tradeflow="imports" if tradeflow=="Imports"
replace tradeflow="reimports" if tradeflow=="Re-imports"

replace reporter="united kingdom" if reporter=="United Kingdom"

replace partner = lower(partner)

gen mdate = ym(floor(period/100), mod(period, 100)) 
gen period_new = mdate - ym(2008, 12)

format %100.0g tradevalueus // um e's wegzubekommen

graph twoway line tradevalueus period_new if tradeflow=="exports" & partner=="belgium", c(L) //

still gives me a strange graph..

[Nick Cox]

The strangeness in the graph in #21 could arise for one or both of two reasons:

1. The observations are out of order. For the graph to be sensible, you need an order in which year varies upwards within blocks of observations for each partner country.

2. There are missing values in the data.

If there are other reasons I can't think of them right now, but that just means what it says.

Either way, further comment requires that you show us the data, namely observations defined by

 
 dataex tradevalueus period_new if tradeflow=="exports" & partner=="belgium"

[Nico Stappen]

I think the data are complete, but probably they are not in order. How could I put them in the right order?

----------------------- copy starting from the next line -----------------------

* Example generated by -dataex-. For more info, type help dataex
clear
input double tradevalueus float period_new
1389925243 129
1109607898 128
1555220225 130
1342127516 126
1486437485 122
1587587248 125
1495376569 121
1703458074 123
1103589216 131
1264351333 127
1283133285 124
1165201243 132
1273365175 143
1241221182 144
1058626805 138
1204877695 141
1149956896 139
 817803249 137
1394906146 142
 998034226 140
1204468225 135
1248488810 134
1283858194 133
 761431299 136
1698244185 156
1890874478 155
1538208377 152
1619116915 150
 850887378 145
1709780733 149
1773939695 146
1716886789 153
1528463529 148
2058386730 147
1835505541 154
1748230297 151
2004987726 158
1435397606 157
end

------------------ copy up to and including the previous line ------------------

Listed 38 out of 796 observations

[Nick Cox]

You see that period_new jumps around, which is why you got the graph you got. As said, you need to sort first on partner country then on the time variable.


help sort

[Nico Stappen]

But is it possible to show a graph which shows the imports and exports of one country (e.g Belgium) over the whole period? I think a bar diagram would be appropriate. So for every period two Bars (1x import, 1x export). X-Axis would be the time. Y-Axis would be the tradevalues. Or am I on the wrong way to get that?

[Nick Cox]

The imports and exports of one country could be shown as an intelligible bar chart if (and I guess only if) you work with totals of each for each year.

[Nico Stappen]

ah okay. Yes the tradevalue data in my data set are the totals of each period, so of each month.
But my data is from January 2019 - April 2022. If it could only be shown in yearly bar diagrams it is only 4 periods, 2 bars each (import and export) would be 8 bars in total. That would be clear but maybe a little bit less expressive .
If I'd show it in monthly bar diagrams that would be 40 month, 2 bars each, so 80 bars. That would be very expressive but hard to show in one diagram I guess.. Do you have an idea how to solve this?

[Nick Cox]

This problem keeps changing. I think I will sign off this thread with a suggestion to look at help egen to find out how to calculate totals over trade partners, which seems to be what your latest version implies.

[Nico Stappen]

I think about how to solve it. But I think with your help my graph now looks good.

Is there a way to get another colored line into this graph with the imports?
And how can I change the x-axis the way that the period isn't from 0-40 but from 01/2019 - 04/2022?