Introduction

With this Best Practices article we start a series reflecting upon the newly released ATLAS.ti 8. The newest version of our software comes with a completely new design and a set of new and improved functions and tools. These changes not only make ATLAS.ti look different from how it looked in earlier versions, but also means that new opportunities for analysis arise.  The ability to now import data from Twitter, Evernote and reference management software clearly opens the doors for analyzing data that could not be analyzed before. But beyond these obvious improvements, ATLAS.ti 8 invites the researcher to rethink how to interact with the data and the kinds of questions to ask.

In this short blog article I will highlight three of what I see as new opportunities for data exploration and analysis that result from the new interface and the new features that come with ATLAS.ti 8.  For a complete description of what is new in this version of the program, please read ATLAS.ti Windows: What’s New.  For more of a complete overview of ATLAS.ti 8, please read the Quick Tour manual.  You may also want to watch the quick tour video tutorials we have made, all of which are available on theATLAS.ti Youtube channel.  In other blog articles in this series we will explore additional features, tools and opportunities for analysis with ATLAS.ti 8.

Working across regions 

Let me start by highlighting the benefits of working across regions, by placing side-by-side and one above the other, as many documents, networks, memos and other tools as you want.  Placing documents side-by-side can be useful as a way of enhancing the capacity for exploration and comparison across documents.  This can be done by segmenting and coding in a quasi-simultaneous way and by exploring relationships between quotations across documents through hyperlinking.  This can allow for more of a holistic understanding of the data than what can be accomplished when working with one document at a time.  See below two documents side-by-side.

Additionally, placing a document next to a memo and a network allows for a nice integration of segmentation and coding, writing, and visualization of linkages. As you read, segment and code the document, you write your reflections on what you are learning, and simultaneously, visualize graphically the linkages you are creating.   See below.

Word Clouds

Word frequencies can be visualized as tables and as word clouds.  I find word clouds particularly useful as a quick and powerful way of seeing what words are more and less frequent in a given document (or a set of documents), a quotation (or a set of quotations), or the quotations linked to a particular code (or the codes belonging to a group).  For example, as you code you may be inclined to quickly see what words are interviewees using to refer to a particular concept (i.e., code), because seeing that can give you insights on what that concept is about from the point of view of study participants.  So, all you do is select the code in the Navigator (left side panel), right-click on it and select ‘Word Cloud’.  See the figure below.

Additionally, you may feel inclined to have a quick view of the words used in given quotation because that could facilitate the understanding of what the interviewee is saying in that segment. Based on that, you could write a comment on the quotation or contribute to one of your memos.  To examine the word cloud of a quotation, select it in the Navigator, right-click on it, and select ‘Word Cloud’.  See the figure below.

Automatic Network Layouts

The newest version of ATLAS.ti comes with twelve network layouts that you can use to graphically represent the linkages between the different elements that make up your analysis project. These are the following:

• Orthogonal
• Orthogonal-Tree
• Circular
• Circular-Single Cycle
• Organic
• Hierarchical
  • Hierarchical Downwards
  • Hierarchical Upwards
  • Hierarchical Left to Right
  • Hierarchical Right to Left
• Tree
• Random

Additionally, the lines linking elements with each other can be routed following four models:

• Rerout Orthogonal
• Polyline
• Rerout Organic
• Rerout Straight

You can learn about the meaning and application of each one of the layouts as well as each one of the line routing models by simply placing the cursor on the corresponding option on the network ribbon.   To make it easy for you, I have included in this article the definition of each layout.  Although some of these definitions use technical language, I have included them here because they are useful to start making sense of what layout is best suited for different kinds of network representations.

Of course, you can always arrange the nodes in the network manually to best reflect the ideas that you want to express.  If you have a taxonomic structure, the Tree or one of the Hierarchical layouts could be the best.  If you have many nodes in a network, I have found that the organic layout works well.

Next, I will show you each one of the twelve layouts using data from a Twitter project. There you will see the following as nodes: hashtags, authors, languages, tweets and the Twitter document.  There is a video in the ATLAS.ti Youtube channel that shows how to work with Twitter data and where some of the network layouts are demonstrated.

Orthogonal

Orthogonal layoouts allow the edges of the graph to run horizontally or vertically, parallel to the coordinate axes of the layout. It produces compact drawings with no overlaps, few crossing and few bends.

Orthogonal-Tree

Same as the standard orthogonal layout, but larger sub-trees are processed using a specialized tree layout algorithm, which is better suited for tree-like structures than the original orthogonal layout style.

Circular

The circular layout places the nodes on a circle, choosing carefully the ordering of the nodes around the circle to reduce crossings and place adjacent nodes close to each other.  It emphasizes group and tree structures within a network. It creates node partitions by analyzing the connectivity structure of the network and arranges the partitions as separate circles.  The circles themselves are arranged in a radial tree layout fashion.  This algorithm suits social network analysis quite well.

Circular-Single Cycle (with the Poly Line rerouting option)

This is similar to the Circular layout, only that sub-groups are not created and all nodes are placed on a single circle. Useful for creating an overview and for shallow hierarchies.

Organic (with the Poly Line rerouting option)

The organic layout style is based on the force-directed layout paradigm.  When calculating a layout, the nodes are considered to be physical objects with mutually repulsive forces like, e.g., protons or electrons.  The connections between nodes also follow the physical analogy and are considered to be springs attached to the pair of nodes.  These springs produce repulsive or attractive forces between their end points if they are too short or too long.  The layout algorithm simulates these physical forces and rearranges the positions in such a way that the sum of the forces emitted by the nodes and the edges reaches a (local) minimum.

Resulting layout often expose the inherent symmetric and clustered structure of a graph, they show a well-balanced distribution of nodes and have few edge crossings.

Radial

In the radial layout style, the nodes of a graph are arranged on concentric circles. The layout calculation starts by conceptually reducing the graph to a tree structure whose root node is taken as the center of all circles.  Each child node in this tree structure is often placed on the next outer circle within the sector of the circle that was reserved by its parent node.  All edges that were initially ignored are re-established and the radii of the circles are calculated taking the sector sizes needed by each whole sub-tree into account.

This layout is well suited for the visualization of directed graphs and tree-like structures.

Hierarchical-Downwards

Prefers to place the nodes downwards from top to bottom along directed links.

Hierarchical-Upwards

Prefers to place nodes upwards from bottom to top along directed links.

Hierarchical-Left to Right

Prefers to place nodes left to right along directed links.

Hierarchical-Right to Left

Prefers to place nodes right to left along directed links.

Tree

The tree layout is designed to arrange directed and undirected trees that have a unique root node.  All children are placed below their parent in relation to the main layout direction.  A child-parent relation in ATLAS.ti is defined via a transitive or asymmetric link. 

Before applying the layout all nodes compromising a strict tree are removed and added after the tree layout by connecting them via curved edges.  Tree layout algorithms are commonly used for visualizing relational data. This layout algorithm starts from the root and recursively assigns coordinates to all tree nodes.  In this manner, leaf nodes will be placed first, while each parent node is placed centered above its children.

Random

Randomly places the node each time this layout is invoked.

Conclusion

It has been interesting to find out how the new and redesigned procedures and tools that came with ATLAS.ti 8 not only allow us to do more and different things, but also present us with new ways of looking at the data. I find that, to some extent, ATLAS.ti 8 allows for a closer interaction with the data than it was possible in earlier versions.  I also find that it allows for an even more fluid integration of the four sub-processes that I consider key in the dialogue that we establish with the data in qualitative analysis: segmentation, writing, coding and visualization of relationships.  The latter is already possible in ATLAS.ti Mac but was missing (at least with the same level of fluidity that is possible now) in the earlier versions of ATLAS.ti Windows.

The very dynamic network tool that has been implemented in ATLAS.ti 8 clearly opens up the possibilities for the representation of linkages and the creation of concepts maps, argumentation maps, taxonomic trees, cognitive maps,  and so forth. As I normally argue in my classes, all of this facilitates one of the central goals in qualitative data analysis: to build a holistic representation of the phenomenon under study.  In future blog articles we will continue examining ATLAS.ti 8 and the new analysis opportunities it brings.

About the Author

Ricardo B. Contreras is an applied anthropologist, researcher and consultant.  He directs the training division of ATLAS.ti Scientific Software GmbH. His research is in migration and community health. You may contact Ricardo by writing to training@atlasti.com.