analisis atlas ti: knowledge workbench.ppt

atlas ti

1. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1

T
h
o
m
a
s

M
u
h
r
Scientific Software Development -
Copyright 2001 Thomas Muhr
This set of 21 PowerPoint transparencies contains information about This set of 21 PowerPoint transparencies contains information about
concepts and use of ATLAS.ti, Please read copyright note on concepts and use of ATLAS.ti, Please read copyright note on
transparency no. 2.transparency no. 2.
ATLASATLAS..titi
The Knowledge WorkbenchThe Knowledge Workbench

2. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1

T
h
o
m
a
s

M
u
h
r
Scientific Software Development -
Copyright 2001 Thomas Muhr
The PowerPoint transparencies included in this package may be The PowerPoint transparencies included in this package may be
used to support your ATLAS.ti workshops, training sessions & used to support your ATLAS.ti workshops, training sessions &
demonstrations.demonstrations.
You may alter the transparencies to fit your needs, but please do not You may alter the transparencies to fit your needs, but please do not
remove original copyright notes. If you have any transparencies remove original copyright notes. If you have any transparencies
either self made or created via modification of the existing sheets we either self made or created via modification of the existing sheets we
will all be happy if you make these available for the public.will all be happy if you make these available for the public.
In no event may the transparencies included in this package be In no event may the transparencies included in this package be
commercially exploited (e.g., sold) either altered or unaltered without commercially exploited (e.g., sold) either altered or unaltered without
prior written permission by the author, Thomas Muhr, Berlin.prior written permission by the author, Thomas Muhr, Berlin.
© Copyright Note© Copyright Note

3. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
ATLASATLAS..titi -- The Knowledge WorkbenchThe Knowledge Workbench
Basics:
QDA & ATLAS.ti
VISE: Visualization, Integration,
Serendipity and Exploration
Users: from Sigmund Freud to Sherlock
Holmes
The main concepts: of Hermeneutic
Units, Families and other species
Strategies: Textual and Conceptual
level
The user interface: Keep focused on
the data
Back to the future: The Paper & Pencil
look & feel
Basic Procedures: Coding,
commenting, retrieving, printing,
preparing,
Beyond Text: Working with
graphics, audio & video materials
Structures: Weaving semantic
networks
Hypertext: What codes can’t do for
you
Retrieval: Using Boolean, Semantic
and Proximity operators
Super Codes: Intensional codes or
frozen hypotheses?
Cooperation: Merging projects
Interfaces: ASCII/ANSI, SPSS,
HTML, PROLOG, WMF, XML
Miscellaneous: Data safety,
memo outsourcing, text
management, setup, capacities
Advanced Topics:

4. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
4 Basic Principles: VISE4 Basic Principles: VISE

VisualizationVisualization

Use adequate tools for handling complexity and stay Use adequate tools for handling complexity and stay
focused on the datafocused on the data

IntegrationIntegration

Bundle all relevant data and interpretations into a unique Bundle all relevant data and interpretations into a unique
project: the “Hermeneutic Unit”project: the “Hermeneutic Unit”

SerendipitySerendipity

Make relevant discoveries without searching...Make relevant discoveries without searching...

ExplorationExploration

Traverse the “interpretative threads” between data, codes, Traverse the “interpretative threads” between data, codes,
and memosand memos

5. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Areas of ApplicationAreas of Application
Criminology
Planning
Applications
Social Sciences
& Humanities
Marketing Research
Libraries & Archives
Urban
Development
Literature
Astronomy
Art
Theology
Medicine
Public Health
Education

6. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Textinterpretation as Text-to-TextTextinterpretation as Text-to-Text
Compile the primary
documents: Texts,
Graphics, Audio, Video
Open up a “Context of Discovery”
to explore the data and add structure
Result: another text,
diagrams, a WWW-
document ?

7. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
A HU’s Abstractional LayersA HU’s Abstractional Layers
Code family
Primary
documents
Quotations
Codes
Super Codes
Families
Networks
contained-in
causesisa
isa
causes
uses
uses
uses
contained-in
indicated-by
indicated-by
indicated-by
supports
Hermeneutic Unit

8. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Graphical Primary DocumentsGraphical Primary Documents
Display comments
for image sections
with a mouseclick
Graphical list of
contents

9. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Industry StandardsIndustry Standards supported supported
by ATLAS.ti 4.2by ATLAS.ti 4.2
S
t
a
n
d
a
r
d
i
z
a
t
i
o
n
Presentation Representation
HTML
XML 1.0
ASCII
WMF
APN
SPSS
RTF
SGML
BMP
TIF, JPG, Kodak PCD, SUN Raster...
Not supported
Exported
Imported
Im- & Export
PCD
ANSI
Will be supported in 5.0
Currently memos and codes

10. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Inter-Application Data ExchangeInter-Application Data Exchange
Text can be dragged from
WinWord or any other text
processors (capable of OLE-
2 drag & drop) into ATLAS/ti.
Text import is also available
via Copy & Paste.
Dropped into a Network View, a
new memo is automatically
created from the text
A mouse click displays the
new memos’ text.

11. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
HTML Code GeneratorHTML Code Generator
The conversion of Hermeneutic Units
into HTML code enables new ways of
structured publishing. Research
teams can quickly exchange ideas
and complete projects world wide.

12. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
The HU Editor The HU Editor
- ATLAS.ti’s main work space- ATLAS.ti’s main work space
Dropdown fields for
Primary Docs,
Quotations,
Codes and Memos
Main menu Main toolbar
Margin area
Splitter bar
to resize panes
Detached
code list
Primary Document
area
Selected Quotation
Context menu

13. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Network EditorNetwork Editor

14. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Hypertext - what codes cannot doHypertext - what codes cannot do
contradicts
Code A
Code B
Q1
Q2
While codes describe similarity of the
coded segments, it is hard to represent
relations (beside the equivalence
relation) between individual segments.
Only direct links (“hyper-links”) between
segments enable the representation of
such local knowledge.
If one would establish a link between the
codes in the example to emulate a hyper
link, we would have to assume that these
codes do not refer to any other
segments, but are used as labels for
individual segments: a clear “misuse” of
codes....
ATLAS/ti supports named links between
data segments.
supports
Q3

15. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
The QueryToolThe QueryTool
The QueryTool retrieves data segments by their attached codes using
Boolean, proximity and semantic operators. Queries are entered in
RPN calculator style.
Boolean
operators
Semantic
operators
Proximity
operators
OR
XOR
AND
NOT
SUB
UP
SIB
WITHIN
ENCLOSES
OVERLAPPED_BY
OVERLAPS
FOLLOWS
PRECEDES
COOCCUR
Term stack
Feedback pane
Results
Stack manipulation
Clear stack
Swap the two
topmost elements
Push - duplicate
topmost element
Recalculate results
Undo last operation
Redo last undone
operation
Create Super Code
Change feedback
display mode
Codes
Families
Textbase selection
Follows/Precedes
distance control

16. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Boolean retrieval is purely set based. Elements are assumed to be
independent. No property of a retrieved segment other than being
coded with codes A,B,..X is taken into account.
Overselectivity: AND (A1, A2, ..., An) fails even with n-1 matching
terms.
Underspecified: OR (A1, A2, ..., An) succeeds with everything from 1
to n matching terms. A segment coded with only one code is treated
equal to one coded with all of them.
Retrieval Methods I - Boolean RetrievalRetrieval Methods I - Boolean Retrieval
A B
not (A or B)
A or B
A and B
not A and B
A xor B
A and not B
Q1
Q2
Q3
Q4
Q5
Document universe: Q1,...,Q5
Query examples:
A -> {Q1, Q2, Q3}
B -> {Q3, Q4}
not A -> {Q4, Q5}
A or B -> {Q1, Q2, Q3, Q4}
A xor B -> {Q1, Q2, Q4}
not (A or B) -> {Q5}
A and not B -> {Q1, Q2}
A and B -> {Q3}
not A and B -> {Q4}

17. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Proximity retrieval takes the spatial relations between the retrieved
elements into account. A segment can overlap, follow, enclose or
simply cooccur with another segment.
The semantics were adapted from Allen’s time logic calculus.
Retrieval Methods II - Proximity RetrievalRetrieval Methods II - Proximity Retrieval
A
B
Q1
Q4
Q5
Q2
Q3
Primary document P1
In addition to the Boolean
conditions described above, the
following proximity relations hold:
B overlaps A -> {Q3, Q4}
A overlapped by B -> {Q1, Q2}
C overlaps B -> {Q5}
A within C -> {Q2}
A overlaps C -> {Q3}
C follows A -> {Q5}
B overlaps C -> {Q3, Q4}
etc.
Note, that proximity operators are
non-commutative:
B op A is not the same as
A op B
Operand input order is significant!
C

18. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Semantic, or thesaurus-based retrieval takes transitive relations
between the terms (codes) into account. Its quality is dependent
upon the quality of the semantic network used.
Retrieval Methods III - Semantic RetrievalRetrieval Methods III - Semantic Retrieval
Q1
Example queries using the semantic operator SUB
on the terminology network below:
sub (Positive Attitude) -> {Q1, Q2, Q3, Q4, Q5}
sub (Negative Attitude} -> {Q6, Q7, Q8}
sub (Attitude) -> {Q1,.., Q8}
While the extension of sub (Positive
Attitude) and or (Love, Kindness) is
identical for the example below
{Q1,..,Q5}, the intension is different.
The former query will - unaltered! -
yield different results with another
subterm of Pos. Attitude. The latter
query will not ecognize this new fact
and has to be reformulated.
Love
Q4
Q5Q2
Q3Q1
Kindness
Positive
Attitude
Q6 Q7 Q8
isa isa
isa
Negative
Attitude
Attitude
isa
Hatred Anger
isa
isa
indicated by
sibling
Document level
Domain level

19. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
QueryTool: Building QueriesQueryTool: Building Queries
Boolean, proximity and semantic operators are combined using
the “click-language” par excellence: the Reverse Polish Notation
(RPN) by Lukasiewicz
1
.
RPN is a parenthesis-free postfix language: operands first, then the
operators.
The main ingredience of the RPN query processor is the Stack, a
data structure, that is very similar to a pile of plates: It can only be
accessed from the top: new plates are put on the pile, plates can
only be removed from the top.
1
Born 1878 in Lvov (now Ukraine), died 1956 in Dublin, Ireland.
Polish Minister of Education in 1919 and professor at Warsaw University from 1920 to 1939)
Old HP 29C RPN calculator
Example: “All quotations coded with ‘Positive Attitude’ and any
of its sub codes but not with ‘Kindness’”
in formal infix notation: SUB Pos. Attitude AND NOT Kindness
Step:1 2 3 4 5
Enter:Pos. Att.SUB (1) Kindness NOT (1) AND (2)
Stack:Pos. Att.SUB(Pos.Att)Kindness NOT(Kindness) AND(NOT(Kindness), SUB(Pos.Att))
- - SUB(Pos.Att) SUB(Pos.Att) -
Result:{} {Q1,...,Q5}{Q3,Q4,Q5} {Q1,Q2,Q6,Q7,Q8}{Q1,Q2}
Note, how every operator takes (“pops”) its appropriate number of arguments from the stack and “pushes” the
resulting term back on the stack. Every entry, operand or operator generates a result. No “syntactic sugar” is
needed as in “infix” notations (eg. parentheses).
Number of arguments

20. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
The QueryTool: Super CodesThe QueryTool: Super Codes
A well constructed, non-trivial query is often the result of a
considerable amount of work and ways to make a query reusable
are needed: Super Codes.
Super Codes are also an important tool for theory construction as
they capture hypotheses for repeated validation against the data.
Normal codes store direct quotation references, super codes store queries.
Although the visible “clicking behavior” of a super code resembles that of normal
codes, there is a considerable difference of “how” each generates its references:
Query X
Normal codes deliver their
quotation references.
The result changes only by
explicitely assigning new or
removing existing references.
Super codes recalculate the
stored query “when-needed”
and deliver the result. When
any of the conditions of the
query change, the super
codes result list changes as
well - without any changes to
the latter.
Unlike other approaches that store the
“extension” (the result set) of a query, super
codes store the queries’ “intension”.
Super codes are “first class”
objects and can be used in
queries (and in other super
codes).

21. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Team Work - Merging Projects ITeam Work - Merging Projects I
Merging projects is mandatory for the support of
teams working on separate data and/or different
code sets.
A number of stock merge strategies permits
efficient control over the resulting project.
Strategies can be freely adapted to fit specific
needs.
Team A
Team A‘s
combined project
All teams‘
combined project
Team B
Team B‘s
combined project

22. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
Merging Projects II - StrategiesMerging Projects II - Strategies
Merging proceeds as subsequent and repeated merging of partial projects into a
target project.
A merge strategy controls the method of how the different object types (e.g.
primary docs, codes) from the source projects migrate into the target project.
Examples:
A Different data sets, same codes
This strategy supports an economic handling of large primary data in a top-
down approach.
B Same data, different codes
By applying this method, different aspects of a theory can be applied to the
same data sets.
Example (P
i
::= primary documents, C
i
::= codes):
HU
1
{P
1
,..,P
n
} {C
1
,..,C
m
} source project
HU
2
{P
1
,..,P
k
} {C
1
,..,C
m
} target project (before merge)
Target project after the merge:
Strategie A: HU {P
1
,..,P
n
,P
n+1
,..,P
n+k
} {C
1
,..,C
m
}
Strategie B: HU {P
1
,..,P
n=k
} {C
1
,..,C
m
,C
m+1
,..,C
m’
}

23. S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t

-

C
o
p
y
r
ig
h
t

2
0
0
1
T
h
o
m
a
s

M
u
h
r
©

S
c
ie
n
t
if
ic

S
o
f
t
w
a
r
e

D
e
v
e
lo
p
m
e
n
t
,

B
e
r
lin
,

2
0
0
1
A
T
L
A
S
.
t
i
®
What’s newWhat’s new in ATLAS.ti 4.2 in ATLAS.ti 4.2
WYSIWYG - printouts of primary texts plus marginWYSIWYG - printouts of primary texts plus margin

Media - fine-grained segmentation and coding of video and Media - fine-grained segmentation and coding of video and
audio filesaudio files (incl. MP3!) (incl. MP3!)
Improved Margin AreaImproved Margin Area

Networks - vector export to drawing software, WordNetworks - vector export to drawing software, Word
®®
etc. etc.
Wordcruncher - count word occurrences and calculate Wordcruncher - count word occurrences and calculate
type/token ratio. type/token ratio.

New reportsNew reports

Primary Doc Path Mapping ToolPrimary Doc Path Mapping Tool
XML XML - memo - memo and code and code import & exportimport & export