5 ms·

The Future of User Interfaces: Data Visualization

38 points by apievangelist 15y ago

ThomPete 15y ago
We are entering an era where the consequences of our interaction with machines won't exist in a one to one relationship but rather in a one to many and many to one. Attempts to manipulate or browse large amounts of data will require much more from the machines than from the humans controlling it. Any routine initiated by the user will be have to be followed by a very large set of sub-routines in order to provide feedback. It might even have to anticipate what kind of feedback the user want. Visualization is going to be the feedback but I don't believe the interface is going to be the screen.
elliottkember 15y ago
I disagree with this article, and agree with the reviewer. The application is not useful at all, and what could have been an excellent UI, given a little more thought and power, is unfortunately just lipstick. Tracks and artists are grouped in 3D space by name alone, and their arrangement and spacing is meaningless. This makes the app a pretty, but useless skin of a simple playlist. I think the design lends itself to music discovery, not playback. Hooked up to rdio or Spotify, this could be a fun way of exploring new music by proximity. Like this song? Here's 5 other similar artists nearby. To me this speaks volumes about data visualisation. Done incorrectly, it's useless. Useful and simple will hopefully win over beautiful but impractical. You can't start with beautiful and add useful later.
- ENOTTY 15y ago
  To build on your comment (which I think is spot on), we've mostly seen data visualization used to build interfaces designed to discover new things from data. Data visualization (currently) seems to be about letting the brain do what it does best and discern patterns from a bunch of data. But when we get into common tasks that have memorized workflows, visualizations might be too clunky of a way to work at it. For any given task, we all have learned a certain workflow that meets our needs in terms of efficiency, complication, and power. These goals aren't always the same goals as a visualization.
- Silhouette 15y ago
  > To me this speaks volumes about data visualisation. Done incorrectly, it's useless. I think it's worse than that: poor visualisations are counterproductive. Visualisation techniques are practically valuable when they let the user understand something that would not otherwise be clear. For example, a tool might analyse a complex data set and use a visualisation that highlights facts that are relevant to answering a particular question. In this case, I don't see what the visualisation tells us that simple tabular playlists don't. It doesn't appear that any use is being made of the extra visual degrees of freedom offered by the astronomical metaphor, and extra visual degrees of freedom (spacial relationships, colours, sizes, shapes...) are usually what makes a good visualisation technique effective.
  - dmlorenzetti 15y ago
    In this case, I don't see what the visualisation tells us that simple tabular playlists don't. To extend your theme with an example-- with a tabular interface, one could sort by song title, in order to listen to covers of a favorite tune. This star-planet-moon interface apparently locks the viewer into sorting only by artist, then by album. That, even more than the wasted opportunity, makes it actively counterproductive.
rushabh 15y ago
I believe the future user interfaces are going to be minimalistic and non-patronizing (like Hacker News). The interface shown in this article (and similar) have a very high learning curve and are counter productive (generally).
- ThomPete 15y ago
  That really depend on what the goal of the interface is. For quickly finding the song/album you want yes, for exploration not necessarily. It would be a stretch to call HN the future of interfaces unless you refer to the underlying algorithms.
  - rushabh 15y ago
    What I meant was that the trend would be more towards minimalism and simple metaphors rather than 3D realism. I don't say that highly visual interfaces don't have their place. I would say the difference is like the difference between a the UI of a book (minimum visual representation) v/s a movie (highest visual representation). If I have to take an honest pick, and I believe I can be wrong, is that most user interfaces will have more abstraction than specific representation.
markkat 15y ago
I think you can change interface, but you can't fundamentally change our sensory organs or how we process information (at least not with an app). We have a biological bias to certain interfaces.
- CognitiveLens 15y ago
  To some extent you are right, but our biological systems are also highly adaptable to new interfaces - e.g. keyboards, home appliances, cars. Interaction design does need to support user perception of how to interact with a device/system, but these affordances can be combined and developed into highly original and useful interactive applications. I agree with others that Planetary does not address these design challenges effectively, but it's important to recognize the adaptability of our sensory systems to fundamentally 'new' ways of interacting with the world.
  - markkat 15y ago
    >but it's important to recognize the adaptability of our sensory systems to fundamentally 'new' ways of interacting with the world. I'd say in most cases, good design asks us to adapt less, rather than more. Keyboards have finger-sized keys, arranged in a heuristic manner informed by linguistics. The keyboard minimizes effort between the brain and text. I agree that Planetary is not effective, and here's why: Planetary arrangements don't map onto music collections naturally, and they do little to inform the organization. Planetary asks us to learn more, to do less. A better music collection interface IMO, would be one where dimensions were informed by music and song information, not planetary physics.
DevX101 15y ago
>One reviewer of the app on iTunes coolly dismissed Planetary as "visually appealing but useless." ...With all due respect, that critic is missing the point. No, he didn't miss the point. He hit the nail on the head. The purpose of data visualization is to help the user quickly understand complex data. This planetary visualization just adds another layer for my brain to parse while giving me nothing in return.

38 points by apievangelist 15y ago

We are entering an era where the consequences of our interaction with machines won't exist in a one to one relationship but rather in a one to many and many to one. Attempts to manipulate or browse large amounts of data will require much more from the machines than from the humans controlling it. Any routine initiated by the user will be have to be followed by a very large set of sub-routines in order to provide feedback. It might even have to anticipate what kind of feedback the user want. Visualization is going to be the feedback but I don't believe the interface is going to be the screen.

elliottkember 15y ago

I disagree with this article, and agree with the reviewer. The application is not useful at all, and what could have been an excellent UI, given a little more thought and power, is unfortunately just lipstick. Tracks and artists are grouped in 3D space by name alone, and their arrangement and spacing is meaningless. This makes the app a pretty, but useless skin of a simple playlist. I think the design lends itself to music discovery, not playback. Hooked up to rdio or Spotify, this could be a fun way of exploring new music by proximity. Like this song? Here's 5 other similar artists nearby. To me this speaks volumes about data visualisation. Done incorrectly, it's useless. Useful and simple will hopefully win over beautiful but impractical. You can't start with beautiful and add useful later.

ENOTTY 15y ago
To build on your comment (which I think is spot on), we've mostly seen data visualization used to build interfaces designed to discover new things from data. Data visualization (currently) seems to be about letting the brain do what it does best and discern patterns from a bunch of data. But when we get into common tasks that have memorized workflows, visualizations might be too clunky of a way to work at it. For any given task, we all have learned a certain workflow that meets our needs in terms of efficiency, complication, and power. These goals aren't always the same goals as a visualization.
Silhouette 15y ago
> To me this speaks volumes about data visualisation. Done incorrectly, it's useless. I think it's worse than that: poor visualisations are counterproductive. Visualisation techniques are practically valuable when they let the user understand something that would not otherwise be clear. For example, a tool might analyse a complex data set and use a visualisation that highlights facts that are relevant to answering a particular question. In this case, I don't see what the visualisation tells us that simple tabular playlists don't. It doesn't appear that any use is being made of the extra visual degrees of freedom offered by the astronomical metaphor, and extra visual degrees of freedom (spacial relationships, colours, sizes, shapes...) are usually what makes a good visualisation technique effective.
- dmlorenzetti 15y ago
  In this case, I don't see what the visualisation tells us that simple tabular playlists don't. To extend your theme with an example-- with a tabular interface, one could sort by song title, in order to listen to covers of a favorite tune. This star-planet-moon interface apparently locks the viewer into sorting only by artist, then by album. That, even more than the wasted opportunity, makes it actively counterproductive.

rushabh 15y ago

I believe the future user interfaces are going to be minimalistic and non-patronizing (like Hacker News). The interface shown in this article (and similar) have a very high learning curve and are counter productive (generally).

ThomPete 15y ago
That really depend on what the goal of the interface is. For quickly finding the song/album you want yes, for exploration not necessarily. It would be a stretch to call HN the future of interfaces unless you refer to the underlying algorithms.
- rushabh 15y ago
  What I meant was that the trend would be more towards minimalism and simple metaphors rather than 3D realism. I don't say that highly visual interfaces don't have their place. I would say the difference is like the difference between a the UI of a book (minimum visual representation) v/s a movie (highest visual representation). If I have to take an honest pick, and I believe I can be wrong, is that most user interfaces will have more abstraction than specific representation.

markkat 15y ago

I think you can change interface, but you can't fundamentally change our sensory organs or how we process information (at least not with an app). We have a biological bias to certain interfaces.

CognitiveLens 15y ago
To some extent you are right, but our biological systems are also highly adaptable to new interfaces - e.g. keyboards, home appliances, cars. Interaction design does need to support user perception of how to interact with a device/system, but these affordances can be combined and developed into highly original and useful interactive applications. I agree with others that Planetary does not address these design challenges effectively, but it's important to recognize the adaptability of our sensory systems to fundamentally 'new' ways of interacting with the world.
- markkat 15y ago
  >but it's important to recognize the adaptability of our sensory systems to fundamentally 'new' ways of interacting with the world. I'd say in most cases, good design asks us to adapt less, rather than more. Keyboards have finger-sized keys, arranged in a heuristic manner informed by linguistics. The keyboard minimizes effort between the brain and text. I agree that Planetary is not effective, and here's why: Planetary arrangements don't map onto music collections naturally, and they do little to inform the organization. Planetary asks us to learn more, to do less. A better music collection interface IMO, would be one where dimensions were informed by music and song information, not planetary physics.

DevX101 15y ago

>One reviewer of the app on iTunes coolly dismissed Planetary as "visually appealing but useless." ...With all due respect, that critic is missing the point. No, he didn't miss the point. He hit the nail on the head. The purpose of data visualization is to help the user quickly understand complex data. This planetary visualization just adds another layer for my brain to parse while giving me nothing in return.