[MUSIC] Welcome back. Your data visualization themselves are only part of the picture. They exist in a format, such as a highly interactive application interface or an infographic poster. Whether it's a dashboard for an operations center or a fun facts graphic being shared on Facebook, the context of your visualization will have an important impact on your design choices. As we have seen in earlier lessons, the level of interactivity can depend on whether the primary purpose is to explain the already known or uncover meaningful and important unknowns. In this lesson you will need to consider the context, potential workflows and interaction design implications of static versus interactive visualizations. Let's have a look. Hi, when designing your visualizations you will need to consider the context and potential workflows in which your designs will be used. If a visualization is static, such as an infographic poster, then it's essential to think very carefully about what data is and it's not being displayed. Why? Well, because there's no way to adjust the view of the final results. However, with interactive visualizations and controls like filtering and zooming you have a lot more option. Of course increasing the level of interactivity demands more from users then a static chart. That is you are requiring people to make their own decisions about what data to look at rather than presenting them with predetermined views. In many cases this increase level of interactivity and user autonomy is essential for the visualization to serve its purpose and meet the users goals. Let's explore interactivity for visualizations a bit more. Large data sets containing complex relationships often generate visualizations that are a total mess. For example, network diagrams can quickly turn into confusing tangles of lines and shapes that are sometimes jokingly referred to as hairballs or balls of spaghetti. This may not be a problem though with the inclusion of filtering and zooming controls that can untangle and make these messes more manageable. Within our activity, you can begin with a big picture view drilled down to a lower level detail and then all within the same workflow interface go somewhere else within the same visualization. Data visualization pioneer, Ben Shneiderman, has a saying or what's sometimes referred to as a mantra for this kind of interaction and data visualization. Overview first, zoom and filter and then details on demand. Now this overview first, zoom and filter process can be iterative so a user may dive in and zoom out of different levels of detail and resolution in the course of their analytic work. Data visualization researcher Colin Ware, has proposed an interesting way of thinking about interactive visualizations in terms of three interlocking feedback loops. Here's a rough, paraphrased description of these loops and some design options and pitfalls that I've seen in the course of my own work. The first is direct manipulation of graphical objects. The second is exploration and navigation. And the third is problem solving and question generation. In your design work, it's good to consider making the most of each of these loops and how they interrelate with each other in the visualization. Let's look at each a little more closely. Direct manipulation of graphical objects is the ability to interact with various visual elements representing data within the visualization. The visual elements on the screen can actively serve many roles for example, by hovering on or rolling over point of interest in a timeline for example. It can provide additional transient, detailed information about that spot. Visual elements can also act as input mechanisms as well as enhanced display items. One of the examples of this is selecting a node in a network diagram to tag and save or to use as the focus in pivot point for a new investigation. These kinds of interactions are commonly done with clicks, right clicks, and small transient menus. While direct manipulation of data is almost a necessity for many purposes there are also some pitfalls which I've regularly seen. Being able to interact with visual elements should be in the service of getting insights and not testing the hand eye dexterity of your users. In some visualizations such stacked column charts and bubble graphs for example, it can be extremely difficult to precisely select and interact with a data point of interest because it's crowded out or covered over by nearby visual elements. Sometimes the issue is that the visual element of interest presents only a very small or ambiguous selectable target area. Techniques such as outlining a particular selected area or highlighting it on rollover can be helpful to address this, as well as the ability to zoom in and out of a particular spot. Exploration and navigation includes the ability for people to easily and effectively find their way through a visual information space. That is, does the interface design help them enter, orient themselves, and get where they want to go, or discover useful pathways ahead and anticipate it? What is the organization of the navigation and the different areas of the screen that people see and use? One issue here is that it's important for people to navigate through an interactive visualization and to keep context from the starting point and know what the options are available as they move through different pathways. Or if they want to go back to a previous point in an analytic process. Things like bread crumbs that show a pathway a user has taken is one basic way to help give users context as they work through a workflow. Problem solving and question generation are two of the primary and ultimate purposes of data visualizations. Do the individual visual elements along with the navigation, help or hinder the ability to problem solve or answer questions that come up along the way. This is a deep topic that extends beyond this lesson. But one of the keys, and this may not surprise you to hear from me, is to know and understand the needs and goals of your intended audiences. If possible get a problem they need to solve. See how they do it currently. Map out a workflow for this process. And then storyboard with visualization design sketches to see how those visualizations might work in this process. We've only scratched the surface of interactive visualizations. You can see there's a lot to this fascinating topic. When you are doing your work think about the different levels of interaction that you should consider, direct manipulation of data representations, navigation and problem solving question generation. Through good design choices, you can enable people to immerse themselves in a process like question and answering sometimes called dynamic querying and focus on gaining insights rather than focusing on distraction in interface or visualization elements. Recall the idea of your work as a data visualization designer as being akin to that of an architect. Just as buildings are designed to provide structures for people to move through, interactive visualizations enable people to navigate or flow through large data sets. Building floor plans and layouts that are shaped based on the needs of people in it and the purposes of the structure is essential. The visual components and information architecture are blueprints of your visualizations, need to support the ability of people to walk through a process of discovery. Thanks for listening.
