Information Visualization (InfoVis) is an exciting field to watch grow and expand into ever new areas. Last year brought some interesting developments that point towards changes in how we do and see visualization. What does 2010 hold in store? Here is a look back and some ideas where we’re heading.

2009: What Was

2009年可视化领域最重要的事是数据获取

Until May 2009, the biggest issue in visualization was access to good data. There were a few standard data sets, and when working with companies we would sometimes get some of their data (which we were not able to share, of course); but comparing visualizations without interesting test data was difficult. Data.gov(and similar websites in other countries) changed all of that. Suddenly, there is more data than we really know what to do with, and more is coming. There are some issues with inflated numbers of data sets (e.g., toxic release data has separate data sets for each state and year), but it’s still a fantastic service and points into the right direction. The recent release of climate data (even if for the wrong reasons) also provided us with some interesting material to dig into. In my opinion, the availability of large amounts of new data was the biggest thing to happen for InfoVis in 2009.

With great data comes great responsibility. We need to figure out ways to do what InfoVis for the Masses has promised for a few years now: easy access to all the relevant data for everybody. Now that we finally have the data, we need to build more and better visualizations. A key event in that direction was the release of Protovis. It is billed as an easy way to build visualizations, and it is very powerful; but the key is that it runs directly in the browser, without the need for plugins. The only issue at the moment is that Internet Explorer does not support the canvas element that Protovis needs, but there are libraries that emulate that using the equivalent IE mechanism. Once this is solved, Protovis will be the way to go for web-based visualization of open data.

What makes Protovis (and also Tableau Public, see below) more interesting thanMany Eyes and other visualization-as-a-service websites so far is the fact that you can create new visualizations for your data, and that you can combine and link several views. The up-front effort to build a visualization is certainly higher, but it also provides a lot more possibilities.

Protovis was presented at InfoVis/VisWeek 2009, and there were also two other papers that will prove to have a lasting effect: Tamara Munzner’s A Nested Model for Visualization Design and Validation and Chris Weaver’s Conjunctive Visual Forms (PDF). While very different, the two papers were similar in one respect: they were theory papers. There haven’t been any theoretical papers in InfoVis recently, and I hope that this is a sign that this type of paper is becoming more acceptable again. The field certainly needs more of the digestion and deep thinking these two examples showed.

An event that involved yours truly was the release of the Parallel Sets program in early June. The tool is finding quite a few users, and we were able to submit an entry to the Discovery Exhibition (and win a prize there). While progress with updates hasn’t been quite as fast as hoped, we have a lot more features planned and hope to release an updated version soon. What is more, we hope to inspire more InfoVis researchers to release their programs to get more tools into people’s hands.

2010: What Will Be

Looking forward, what can we expect from the new year? One product that will have a big impact is Tableau Public. Currently in beta, it will be released in early 2010. It’s a fully functioning version of Tableau Desktop, but it can only save its data to Tableau’s public server (and there are limits on dataset size and data sources). Unlike Protovis, Tableau Public is server-based, with an embedding mechanism that only requires JavaScript. Like Protovis, Tableau offers the possibility to combine different visualizations into a view, with linking between them. The setup is much easier though, and requires no programming.

I will be writing more about JavaScript-based in-browser visualization soon, because I believe that this is the way to go for anything that is meant to be presented to the general public. JavaScript works everywhere, it’s fast and flexible. There are already a number of visualization frameworks with varying levels of sophistication out there, though Protovis and Tableau Public are by far the most interesting. Together with the availability of lots of data, they will enable InfoVis for the Masses to make a big leap forward in 2010, and finally really reach the masses.

Data.gov et al. are a great start, but a lot more is needed. A lot of the data on data.gov is not of much interest to end users right now, and there hasn’t been a lot of new material in a while. That will change this year. Obama’s Open Government Directive has the potential of opening the floodgates on tons of data that is currently only kept secret because nobody is bothering to publish it. The promise of open data has also been so clear and so loud that a lot more action has to follow for people to really believe it. I am convinced that we will finally see the hundreds of thousands of data sets on data.gov that were rumored, and that many of those data sets will be of much greater interest than what we have seen so far.

In addition to the practical visualization uses, 2010 might be the year of visualization theory. While our field is certainly an applied one, we still need a much deeper understanding of how it works and how to build better tools. There is some existing work, but much of that is old (Bertin’s work was published in the 1960s, Mackinlay’s almost 25 years ago, Shneiderman’s 13 years ago, Chi’s taxonomy almost ten years ago). The field is progressing and we are developing new tools that do not always fit the old molds. We are also gaining a better understanding of how things work, and we are seeing interesting new concepts from other fields. So an update of our theoretical foundations is really overdue now, and this year will hopefully be when it happens.

Another trend that seems to be obvious is a continued interest in visualizing data from bioinformatics. There were seven papers at InfoVis and VAST last year (including the InfoVis Best Paper and an Honorable Mention), but there are still a lot of open problems and terribly unusable tools in bioinformatics. We can have a big impact there, and maybe influence the thinking there to become a lot more visual.

Beyond 2010: What Is to Come

In the longer term, the key will be to engage people’s visual thinking skills. As long as visualization is mostly representation, it adds too little to what people can already do. Once we are able to not just show them the data, but stimulate them to think about (and with) it, they will realize the value of visualization. This is true for general audiences as well as very specialized uses: we need to move from visualization to visual analysis, and from representation to problem-solving.

Engaging people does not require interaction, but it will play a bigger role in the future. Interaction in visualization has been stagnant for a while, but I predict that a new class of interactions will be developed before too long. There are some obvious things (like touch interaction), but there has to be more than pointing at objects on the screen. How can we better support efficient visual work with complex data? How can we select in many dimensions at once? How can we specify complex filters and criteria without lots of tedious steps?

Bioinformatics will keep us busy for some time. There is a lot of data there, and more is generated all the time. There’s almost any type of data you want to work with: unstructured, graphs, trees, temporal, etc. And there are lots of issues with uncertainty, data quality, data size, etc. that require new ideas and solutions. While I doubt that it will happen, we could do a lot worse than become a part of bioinformatics. The funding opportunities alone are mind-blowing.

Another very promising area is journalism. As it is increasingly based on data, InfoVis will play a bigger role in the analysis that leads to a story, as well as its presentation. At the moment, charts are only done to support a story and to give people something to play with. But why not make the visualization the central element, and only provide textual explanations as needed? To make that feasible, much better tools for creating visualizations will be needed that do not require programming but still enable the user to create new types of highly specific visualizations.

It’s safe to say that we will not run out of interesting work anytime soon. We will need to break out of the ivory tower more if we want to stay relevant, though. Given the visual nature of our field, that should not be too difficult, though.

最常见的数据展示方式是一个名词,后面跟一个数字,比如显示病人的血糖含量:

血糖:128

像上面把一个数字放在恰当的地方,这个数字就有个额外的含义.但现代的医学在测量病人的血糖是通常要和之前的测量记录进行比较.下面用一条线表示最近80次的测量结果:

血糖:128

在没有放大的情况下,我们无法精确得知每一点的数据,但我们至少可以知道个大概,我们还可以用颜色把数据点和数值”绑定”起来

血糖:128

1.安装

首先要下载一个名为Open Flash Chart .zip 的文件。

解压zip文件，里面有一个version-2文件夹，里面有很多东西，不过只有一个swf文件是必要的，复制这个文件夹里的open-flash-chart.swf 文件到你的网站目录下（其实任意地方都可以啦）。

2. 编辑HTML文件

First create a new HTML file that will show a simple ‘hello world!’ page. Copy the following into a new file, save it as ‘chart.html‘ in the root of your website.

先新建一个html文件，比如叫“chart.html”,把下面的代码复制进去：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

<html>
<head>
 
 
 
<p>Hello World</p>
 
 
<object classid="clsid:d27cdb6e-ae6d-11cf-96b8-444553540000"
        codebase="http://fpdownload.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=8,0,0,0"
        width="500"
		height="250" id="graph-2" align="middle">
 
	<param name="allowScriptAccess" value="sameDomain" />
	<param name="movie" value="open-flash-chart.swf" />
	<param name="quality" value="high" />
	<embed src="open-flash-chart.swf" quality="high" bgcolor="#FFFFFF" width="500" height="250" name="open-flash-chart" align="middle" allowScriptAccess="sameDomain" type="application/x-shockwave-flash" pluginspage="http://www.macromedia.com/go/getflashplayer" />
</object>

这个html的作用就是把swf嵌入到网页里。

现在用浏览器打开这个html文件，你可以看到一个hello world的页面，不过没有图表，为啥？因为还没加数据呗^_^

3添加数据

大家都知道，图表的显示是基于数据的，ofc当然也需要数据，ofc使用的是一种叫做JSON的数据格式（读音像节丧），这种格式类似AS里的object，就是key：value这种，下面是一个JSON数据的示例：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

{
  "title":{
    "text":  "Many data lines",
    "style": "{font-size: 20px; color:#0000ff; font-family: Verdana; text-align: center;}"
  },
 
  "y_legend":{
    "text": "Open Flash Chart",
    "style": "{color: #736AFF; font-size: 12px;}"
  },
 
  "elements":[
    {
      "type":      "bar",
      "alpha":     0.5,
      "colour":    "#9933CC",
      "text":      "Page views",
      "font-size": 10,
      "values" :   [9,6,7,9,5,7,6,9,7]
    },
    {
      "type":      "bar",
      "alpha":     0.5,
      "colour":    "#CC9933",
      "text":      "Page views 2",
      "font-size": 10,
      "values" :   [6,7,9,5,7,6,9,7,3]
    }
  ],
 
  "x_axis":{
    "stroke":1,
    "tick_height":10,
    "colour":"#d000d0",
    "grid_colour":"#00ff00",
    "labels": ["January","February","March","April","May","June","July","August","Spetember"]
   },
 
  "y_axis":{
    "stroke":      4,
    "tick_length": 3,
    "colour":      "#d000d0",
    "grid_colour": "#00ff00",
    "offset":      0,
    "max":         20
  }
}

可以新建一个文本文档，把上面的内容复制进去，然后重命名为data.json。

现在数据也有了，那怎么把数据放到网页的flash图表里去呢，有一个最简单的方法，有两个步骤：

• 把data.json和chart.html放在同级目录下。
• 在URL后面加上”?ofc=data.json”,比如：www.mojojo.org/chart.html?ofc=data.json。

现在看看，是不是可以看到了呢？

原文：http://teethgrinder.co.uk/open-flash-chart-2/tutorial.php