Introduction to the Participatory Workshop Series & Stata ... · Introduction to the Participatory...

Introduction to the Participatory Workshop Series& Stata graphics sneak peek

Gustavo CaballeroEconomics Graduate Association

2nd of July 2014

EGA (U. of Calgary) Introduction to the PWS & Stata graphics 02/07/2014 1 / 20

The PWS Stata Graphics sneak peek

Table of Contents

1 The PWS

2 Stata Graphics sneak peek



The Participatory Workshops seriesHow we think writing a document is:



The Participatory Workshops seriesHow writing a document really is:



PWS - Why should we have them

Learning how to give a publication quality to our documents takes alot of time; time that we could spend in improving the content of ourdocument.

Don’t worry. Eventually we all figure out a way.I This way can be oftentimes complicated but sometimes we also find a

significant number of very good “tricks”.I Even if you have found a trick, the more ways you know how to do

something (graphics, tables, etc) the better. Therefore we could all winif we shared our tricks.

Students get increasingly busy and they don’t have a space for this,the tricks they learned leave the department with them and theprocess start again every generation.

This Series of Workshops is that space.



PWS - The plan

The idea is all sharing their ways—tricks—to make tables, graphics,etc. (subjects are selected in advance). Therefore the Participatorypart of the series title.

Though we will all share, it is better if we have a Workshop Leaderand it can be you.

So far, the members of the executive committee have thought on thefollowing subjects:

I Graphics in Stata; if someone knows R or Mathematica there could beone in those programs.

I Regression output in Stata for Word/LaTeX.I Descriptive tables in Stata.I An introduction to LaTeX (TeXMaker).I Managing references in Mendeley.

What else??



Table of Contents

1 The PWS

2 Stata Graphics sneak peek



The idea of a graphic (figure) is to transmit a bigamount of information in a condensed AND clear way.



Stata Graphics: a pie chart example.Excel 3D. Really??



Stata Graphics: a pie chart example.Excel 3D. Really??

Despite being very popular outside the academia, 3D graphics are notappropriate for an academic document as it goes against clarity.

I Perspective in 3D graphics distorts proportions. In this case, forexample it makes the categories depicted in the lower part of the pieseem bigger when compared to categories in the upper part.

I Which one is bigger, orange or blue?

In the Workshop someone said that we could include thenumbers/percentages, I will come back to that in the end.



Stata GraphicsSimple graphic. What is this?

sales mktingrsrch dvpt

graph pie sales mkting rsrch dvpt



A graphic must be self-contained meaning that it should have all theinformation needed to understand what it is presenting to the reader.

Many of us are wrong when we believe that including an explanationin the written part is enough.

I In this case, it would be believing that because we write aboutmarketing or research then the reader can understand what is mktingor rsrch.

There is no additional information as to what the slides represent:number of employees?, costs? or revenue? of which company?,sector? or country?

All variables should be given a variable label before making anytables/graphics.

The same is true of value labels when you have categorical variables,for example.



Stata GraphicsNo legend, title, subtitle, labels and notes

Sales

Marketing

Research

Development

Source: 2002 Financial Report (fictional data)

2002Expenditures, XYZ Corp.

label var sales "Sales"

label var mkting "Marketing"

label var rsrch "Research"

label var dvpt "Development"

graph pie sales mkting rsrch dvpt, ///

plabel( all name, size(* 1.5) color(white)) ///

legend(off) ///

plotregion(lstyle(none)) ///

title("Expenditures, XYZ Corp.") ///

subtitle("2002") ///

note("Source: 2002 Financial Report (fictional

data)")



This graphic is much better. We do not explain the commands forthe changes—as these tend to be specific to the type of graphiccommand— but we say the objects included.

We included:I A title This title should be very short though sufficient to explain what

the graphic is; now we know the graphic is representing expenditures.I A subtitle may be included or not depending whether we want to be

more specific, in this case the graphic refers to the year 2002.I Labels as explained before. Note that we got rid of the legend. If

possible do that as legends tend to take a lot of space.I Notes it is often recommended to explain the source of your data as in

cases someone may have some questions or would like to reproduce thetable.



Stata GraphicsNo background color and grayscale.

Sales

Marketing

Research

Development

Source: 2002 Financial Report (fictional data)

2002Expenditures, XYZ Corp.

set scheme s2mono

graph pie sales mkting rsrch dvpt, ///

plabel( all name, size(*1.5) color(white)) ///

legend(off) ///

title("Expenditures, XYZ Corp.") ///

subtitle("2002") ///

note("Source: 2002 Financial Report (fictional

data)") ///

graphregion(fcolor(white) lcolor(white))



Finally we want to get rid of that awful background that Stataincludes with the option graphregion(fcolor(white)lcolor(white)). Remember that the graphic should be as clear asit can and that background is nothing but distracting.

I You must think a little, or try placing it in different parts of thecommand, if you are using the by() option or the command graph

combine.

Though the graphics with colors do seem much nicer, the truth isthat most of readers will print your paper in black and white. In thiscase, colors as the red and blue in slide 13 can be very similar onceprinted. Therefore it is recommended using a grayscale.

EXTRA: it is recommended to place the captions on graphics below,and not above the graphic (as done in tables). This is just as a resultof how people approach a graphic object.



Stata GraphicsExcel can come close but not the same. (Hard to include notes)



The most important point!

As we said in the beginning, graphics should only be used whenyou are summarizing a lot of data.Coming back to the point made on including the numbers in the3D graphic:

I What is presented in this Pie Chart could have been easily discussed inone sentence.

I That is why you do not find this type of graphics in academicdocuments.

I (Don’t use Pie Charts)

Use graphics/figures only when needed.



Participants comments:Included after the Workshop.

When making your graphics, stay with only one program if possible,in this way all your graphics will have the same “look”.

Do not make adjustments to your graphics manually. If you need tochange something then you need to go over all the steps again andyou may forget how you did something.

I Also, you may even don’t remember how you did a certain figure and ifsomeone ask you how you did it, e.g. supervisor or referee), you will bein a very bad position.

When using scatter-plots use hollow dots instead of full dots.

Some students would like to see workshops in graphics andprogramming in Matlab. A leader is needed for these ones.



Thank you.


Introduction to the Participatory Workshop Series & Stata ... · Introduction to the Participatory...

Documents

Transcript of Introduction to the Participatory Workshop Series & Stata ... · Introduction to the Participatory...