Introduction to Stata 13 Graphics - Office of Population Research

Introduction to Stata 13 Graphics
Dawn Koffman
Office of Population Research
Princeton University
September 2013
1
Stata 13 Graphics
Pros:
Many graph types and plot types provided
Multiple plot types may be overlaid
Can easily change overall look of graphs
Same options available for most types of graphs
Very flexible
Cons:
Sometimes slow
Large syntax: 680 page graphics manual!
Stata 13 Graphics Manual is only available on-line:
Help -> PDF Documentation -> [G] Graphics
Stata Graphics References:
http://data.princeton.edu/stata/Graphics.html, by German Rodriguez
A Visual Guide To Stata Graphics, Third Edition, by Michael Mitchell
Stata 13 Graphics Manual (may want to start with “graph intro”)
2
Stata Graphics Syntax
graph <graphtype>
graph bar
graph twoway <plottype>
graph twoway scatter
graph twoway line
graph twoway lfit
graph twoway lfitci
graphs commands may have options
some options have suboptions or a list of options
graph twoway scatter var1 var2, xlabel(30(10)100, labsize(small))
appearance of graph defined by graph elements:
data - marker symbols, lines
elements within plot region – text, marker labels, line labels
elements outside plot region – titles, legend, notes, axis labels, tick marks, axis titles
size and shape of plot region and entire graph
3
/* OR */
50
twoway line le year
70
graph twoway line le year
life expectancy
60
sysuse uslifeexp.dta, clear
80
Stata Graphics Syntax: A Simple Example
40
/* OR */
line le year
1900
1920
1940
1960
1980
2000
Year
4
80
40
50
line le year, scheme(s1mono)
life expectancy
60
70
Using Schemes
1900
1920
1940
1960
1980
2000
Year
80
line le year, scheme(economist)
/* to see list of
scheme names:
graph query, schemes
to change default scheme:
set scheme schemename
*/
60
life expectancy
70
50
1900
1920
1940
1960
1980
40
2000
Year
5
Multiple Dependent Variables
30
40
50
60
70
80
line le_wmale le_wfemale le_bmale le_bfemale year
1900
1920
1940
1960
1980
2000
Year
Life expectancy, white males
Life expectancy, black males
Life expectancy, white females
Life expectancy, black females
6
Adding Text
40
50
60
70
80
line le_wmale le_wfemale le_bmale le_bfemale year ///
, text(32 1920 “{bf:1918} {it:Influenza} Pandemic", place(3))
30
1918 Influenza Pandemic
1900
1920
1940
1960
1980
2000
Year
Life expectancy, white males
Life expectancy, black males
Life expectancy, white females
Life expectancy, black females
7
Overlaying Two-Way Plot Types
scatter le year if year >= 1950 || lfit le year if
year >= 1950
twoway ///
(scatter le year if year >= 1950) ///
(lfit
le year if year >= 1950)
74
72
/* OR */
70
scatter ///
le year if year >= 1950 ///
|| lfit le year if year >= 1950
76
/* OR */
#delimit
twoway
(scatter
(lfit
#delimit
;
68
/* OR */
1950
le year if year >= 1950)
le year if year >= 1950);
cr
1960
1970
life expectancy
Year
1980
1990
2000
Fitted values
8
Overlaying Two-Way Plot Types
>=
>=
<
>=
1925
///
1925 & ///
1950
///
1950
75
scatter le year if year
|| lfit le year if year
year
|| lfit le year if year
65
60
///
>= 1925) ///
>= 1925 & ///
< 1950) ///
>= 1950)
55
twoway
(scatter le year if year
(lfit
le year if year
year
(lfit
le year if year
70
/* OR */
/* OR */
1920
#delimit ;
scatter le year if year >= 1925
|| lfit le year if year >= 1925 & year < 1950
|| lfit le year if year >= 1950;
#delimit cr
1940
1960
Year
life expectancy
Fitted values
1980
2000
Fitted values
9
Overlaying Two-Way Plot Types
#delimit ;
scatter le_male le_female year if year >= 1950
|| lfit le_male
year if year >= 1950
75
70
65
#delimit cr
80
|| lfit le_female year if year >= 1950;
1950
1960
1970
1980
1990
2000
Year
Life expectancy, males
Fitted values
Life expectancy, females
Fitted values
10
Adding a Title and Removing the Legend
75
80
#delimit ;
scatter le_male le_female year if year >= 1950
|| lfit le_male
year if year >= 1950
|| lfit le_female year if year >= 1950
,title("US Male and Female Life Expectancy, 1950-2000")
text(75 1978 "Female", place(3))
text(68 1978 "Male", place(3))
legend(off);
US Male and Female Life Expectancy, 1950-2000
#delimit cr
70
Female
65
Male
1950
1960
1970
1980
Year
1990
2000
11
Showing Confidence Intervals, Labelling Axes, Modifying Legend
sysuse lifeexp.dta, clear
#delimit ;
twoway
(lfitci lexp safewater if region == 2) /* North America */
(scatter lexp safewater if region == 2)
,title("Life expectancy at birth by access to safe water, 1998")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(ring(0) pos(5) order(2 "Linear fit" 1 "95% CI"));
#delimit cr
Life expectancy at birth
60
70
80
Life expectancy at birth by access to safe water, 1998
50
Linear fit
20
40
60
80
Percent of population with access to safe water
95% CI
100
12
Markers Labels and Subtitles
twoway
(lfitci lexp safewater if region == 2) /* North America */
(scatter lexp safewater if region == 2, mlabel(country))
,title("Life expectancy at birth by access to safe water, 1998")
subtitle("North America")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(ring(0) pos(5) order(2 "Linear fit" 1 "95% CI"));
Life expectancy at birth by access to safe water, 1998
80
North America
Canada
CubaPuerto Rico
Jamaica
Life expectancy at birth
60
70
Panama
Trinidad and Tobago
Mexico
Dominican Republic
El Salvador Honduras
Nicaragua
Guatemala
Haiti
50
Linear fit
20
40
60
80
Percent of population with access to safe water
95% CI
100
13
Position of Marker Labels
Life expectancy at birth by access to safe water, 1998
80
North America
Canada
Cuba
Panama
Life expectancy at birth
60
70
Jamaica
Trinidad and Tobago
Dominican Republic
Honduras
El Salvador
Puerto Rico
Mexico
Nicaragua
Guatemala
50
Haiti
Linear fit
generate pos = 12 if country == "Panama"
20
40
60
80
replace pos = 12 if country == "Honduras"
Percent of population with access to safe water
replace pos = 10 if country == "Cuba"
replace pos = 9 if country == "Jamaica"
replace pos = 9 if country == "El Salvador"
replace pos = 9 if country == "Trinidad and Tobago"
replace pos = 9 if country == "Dominican Republic"
#delimit ;
twoway
(lfitci lexp safewater if region == 2) /* North America */
(scatter lexp safewater if region == 2
, mlabel(country) mlabvposition(pos))
,title("Life expectancy at birth by access to safe water, 1998")
subtitle("North America")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(ring(0) pos(5) order(2 "Linear fit" 1 "95% CI"))
plotregion(margin(r+10));
#delimit cr
95% CI
100
14
Position of Marker Labels
#delimit ;
twoway
(scatter lexp safewater if region == 2 | region == 3
,mlabel(country))
,title("Life expectancy at birth by access to safe water, 1998")
subtitle("North and South America")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
plotregion(margin(r+10));
#delimit cr
Life expectancy at birth by access to safe water, 1998
80
North and South America
Life expectancy at birth
60
65
70
75
Canada
CubaPuerto Rico
Chile
Panama
Uruguay
Argentina
Venezuela
Trinidad and Tobago
Mexico
Dominican Republic
Ecuador Colombia
El SalvadorHonduras
Peru
Nicaragua
Brazil
Jamaica
Paraguay
Guatemala
55
Bolivia
Haiti
20
40
60
80
Percent of population with access to safe water
100
15
and
Legend Display
Life expectancy at birth by access to safe water, 1998
80
North and South America
Life expectancy at birth
60
65
70
75
Position of Marker
Labels
Canada
Puerto Rico
Cuba
Jamaica
Chile
Panama
Uruguay
Argentina
Venezuela
Trinidad and Tobago
Mexico
Dominican Republic
Paraguay
Ecuador
Colombia
El Salvador
Honduras
Peru
Nicaragua
Brazil
Guatemala
Bolivia
55
North America
replace pos = 9 if country == "Argentina"
South America
Haiti
replace pos = 9 if country == "Canada"
replace pos = 9 if country == "Cuba"
20
40
60
80
100
replace pos = 9 if country == "Panama"
Percent of population with access to safe water
replace pos = 9 if country == "Venezuela"
replace pos = 9 if country == "Jamaica"
replace pos = 9 if country == "Dominican Republic"
replace pos = 9 if country == "Ecuador"
replace pos = 9 if country == "El Salvador"
replace pos = 12 if country == "Puerto Rico"
#delimit ;
twoway
(scatter lexp safewater if region == 2
,mlabel(country) mlabvposition(pos))
(scatter lexp safewater if region == 3
,mlabel(country) mlabvposition(pos))
,title("Life expectancy at birth by access to safe water, 1998")
subtitle("North and South America")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(ring(0) pos(5) order(1 "North America" 2 "South America") cols(1));
16
#delimit cr
Marker Size and Symbol, Line Color
twoway
(scatter lexp safewater if region == 2
,mlabel(country) mlabvposition(pos) msize(small))
(scatter lexp safewater if region == 3
,mlabel(country) mlabvposition(pos) msize(small) msymbol(circle_hollow))
(lfit lexp safewater if region == 2, clcolor(navy))
(lfit lexp safewater if region == 3, clcolor(maroon))
,title("Life expectancy at birth by access to safe water, 1998")
subtitle("North and South America")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(ring(0) pos(5) cols(1) order(1 "North America" 2 "South America"
3 "North America linear fit" 4 "South America linear fit"));
Life expectancy at birth by access to safe water, 1998
80
North and South America
Canada
Life expectancy at birth
60
65
70
75
Puerto Rico
Cuba
Chile
Panama
Uruguay
Argentina
Venezuela
Trinidad and Tobago
Mexico
Dominican Republic
Paraguay
Ecuador
Colombia
El Salvador
Honduras
Peru
Nicaragua
Brazil
Jamaica
Guatemala
55
Bolivia
Haiti
20
North America
South America
North America linear fit
South America linear fit
40
60
80
Percent of population with access to safe water
100
17
Marker and Marker Label Color, Line Style
twoway
(scatter lexp safewater if region == 2
,mlabel(country) mlabvposition(pos) msize(small) mcolor(black) mlabcolor(black))
(scatter lexp safewater if region == 3
,mlabel(country) mlabvposition(pos) msize(small) mcolor(black) mlabcolor(black)
msymbol(circle_hollow))
(lfit lexp safewater if region == 2, clcolor(black))
(lfit lexp safewater if region == 3, clcolor(black) clpattern(dash))
,title("Life expectancy at birth by access to safe water, 1998", color(black))
subtitle("North and South America")
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(ring(0) pos(5) cols(1) order(1 "North America" 2 "South America"
3 "North America linear fit" 4 "South America linear fit"));
Life expectancy at birth by access to safe water, 1998
80
North and South America
Canada
Life expectancy at birth
60
65
70
75
Puerto Rico
Cuba
Chile
Panama
Uruguay
Argentina
Venezuela
Trinidad and Tobago
Mexico
Dominican Republic
Paraguay
Ecuador
Colombia
El Salvador
Honduras
Peru
Nicaragua
Brazil
Jamaica
Guatemala
55
Bolivia
Haiti
20
North America
South America
North America linear fit
South America linear fit
40
60
80
Percent of population with access to safe water
100
18
By-Graph: Separate Graphs for Each Subset of Data
60
50
Total
60
70
80
S.A.
50
Life expectancy at birth
70
80
#delimit ;
twoway scatter lexp safewater, by(region, total)
,ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe
water");
#delimit cr
Eur & C.Asia
N.A.
20
40
60
80
100 20
40
60
Percent of population with access to safe water
Graphs by Region
80
100
19
By-Graph Options
twoway scatter lexp safewater
,by(region,total style(compact)
title("Life expectancy by access to safe water") note(""))
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water");
Life expectancy by access to safe water
N.A.
S.A.
Total
60
50
80
70
60
50
Life expectancy at birth
70
80
Eur & C.Asia
20
40
60
80
10020
40
60
Percent of population with access to safe water
80
100
20
Axis Scale, Ticks and Labels
twoway scatter lexp safewater
, by(region,total style(compact)
title("Life expectancy by access to safe water") note(""))
xscale(range(20 100))
xtick(20(10)100)
xlabel(30(10)100, labsize(small))
xtitle("Percent of population with access to safe water")
ytitle("Life expectancy at birth")
ylabel(55(5)80, angle(0));
Life expectancy by access to safe water
Eur & C.Asia
N.A.
S.A.
Total
80
75
Life expectancy at birth
70
65
60
55
80
75
70
65
60
55
30
40
50
60
70
80
90
100
30
40
50
60
Percent of population with access to safe water
70
80
90
100
21
Storing Graphs in Memory
twoway
(scatter lexp safewater if region == 2,
mcolor(black) msize(small)
mlabel(country) mlabvposition(pos) mlabcolor(black))
(lfit lexp safewater if region == 2, clcolor(black))
,name(north_america, replace)
subtitle("North America", color(black))
ylabel(,angle(0))
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(off);
North America
80
Canada
Puerto Rico
Cuba
Life expectancy at birth
75
Jamaica
Panama
Trinidad and Tobago
Mexico
Dominican Republic
70
El Salvador
Honduras
65
Nicaragua
Guatemala
60
55
Haiti
20
40
60
80
Percent of population with access to safe water
100
22
Storing Graphs in Memory
twoway
(scatter lexp_sa safewater if region == 3,
mcolor(black) msize(small)
mlabel(country) mlabvposition(pos) mlabcolor(black))
(lfit lexp safewater if region == 3, clcolor(black))
,name(south_america, replace)
subtitle("South America", color(black))
ylabel(, angle(0))
ytitle("Life expectancy at birth")
xtitle("Percent of population with access to safe water")
legend(off);
South America
75
Chile
Uruguay
Life expectancy at birth
Argentina
70
Paraguay
Venezuela
Ecuador
Colombia
Peru
Brazil
65
Bolivia
60
40
50
60
70
80
Percent of population with access to safe water
90
23
Combining Graphs
graph combine north_america south_america
,title("Life expectancy by access to safe water", color(black)) col(1);
Life expectancy by access to safe water
Life expectancy at birth
North America
80
75
Canada
Puerto Rico
Cuba
Panama
Trinidad
Mexico and Tobago
Jamaica
70
El Salvador
Dominican Republic
Nicaragua
Honduras
Guatemala
65
60
55
Haiti
20
40
60
80
Percent of population with access to safe water
100
Life expectancy at birth
South America
75
Chile
Argentina
70
Paraguay
Venezuela
Ecuador
Uruguay
Colombia
Peru
Brazil
65
Bolivia
60
40
50
60
70
Percent of population with access to safe water
80
90
24
Life expectancy by access to safe water
North America
80
Canada
Life expectancy at birth
75
Jamaica
Dominican Republic
70
Puerto Rico
Cuba
Nicaragua
Honduras
graph combine north_america south_america
,title
("Life expectancy by access to safe water",
color(black))
xcommon ycommon
xsize(7) ysize(10.5)
col(1);
Guatemala
60
55
Haiti
20
Graphs
Panama
Trinidad and Tobago
Mexico
El Salvador
65
Combining
40
60
80
Percent of population with access to safe water
100
South America
80
Life expectancy at birth
75
Argentina
70
Paraguay
Venezuela
Ecuador
Chile
Uruguay
Colombia
Peru
Brazil
65
Bolivia
60
55
20
40
60
80
Percent of population with access to safe water
100
25
Saving and Including Stata Graphs
save graph in portable format (format determined by filename extension)
vector formats contain drawing instructions (.wmf .emf .ps .eps .pdf)
resolution independent
work well if graph my be resized
graph export north_america.wmf
raster formats save graph pixel-by-pixel (.png)
use current resolution
work well if including graph on web pages
graph export north_america.png
include "portable-format-graph" in Windows application (Word, Powerpoint):
Insert -> Picture -> From File
26
Using Mata Functions to Add Graphs to Word Document*
create Stata graphs and and use graph export to save graphs in portable format
sysuse uslifeexp
line le year
graph export us_lifeexp_overall.emf, replace
line le_wmale le_wfemale le_bmale le_bfemale year
graph export us_lifeexp_race_gender.emf, replace
use Mata functions to:
create Word document
add Stata graphs
save Word document
mata:
dh = _docx_new()
_docx_image_add(dh, “us_lifeexp_overall.emf”)
_docx_image_add(dh, “us_lifeexp_race_gender.emf”)
rc = _docx_save(dh, “us_lifeexp_graphs.docx”)
end
* New functions available in Stata 13
27