Part I: Entry-level tools

1、Excel

The graphic function of Excel is not powerful, but Excel is an ideal tool for analyzing data. The above picture is a heat map generated by Excel.

Excel, as an entry-level tool, is an ideal tool for quickly analyzing data, and can also create data charts for internal use. However, the range of colors, lines and styles that Excel can choose is limited, which means that it is difficult to make data charts that meet the needs of professional publications and websites with Excel. But as an efficient internal communication tool, Excel should be one of the necessary tools in your treasure chest.

2、CSV/JSON

Although CSV (Comma Separated Values) and JSON(JavaScript Object Annotation) are not real visualization tools, they are common data formats. You must understand their structure and know how to import or export data from these files. All the data visualization tools described below support at least one of CSV and JSON formats.

Part II: On-line data visualization tools.

3、GoogleChartAPI

GoogleChartAPI toolset cancels the static picture function, and only provides dynamic chart tools at present. It can be used in all browsers that support SVG\Canvas and VML, but a big problem with GoogleChart is that charts are generated on the client side, which means that those devices that don't support JavaScript can't be used, can't be used offline or save the results in other formats, so the previous static pictures don't have this problem. Despite the above problems, it is undeniable that GoogleChartAPI is extremely rich in functions. If you don't have special customization requirements or don't conflict with Google's visual style, you can start with GoogleChart.

4、Flot

Flot is an excellent wireframe chart library, which supports all browsers that support canvas (currently mainstream browsers such as Firefox, IE, Chrome, etc.). ).

5. Raphael? l

Raphael L is a JavaScript library for creating charts and graphs. The biggest difference from other libraries is that the output format is limited to SVG and VML. SVG is a vector format, which can be displayed well at any resolution.

6、D3

D3(DataDrivenDocuments) is another JavaScript library that supports SVG rendering. But D3 can provide a large number of complex chart styles besides line charts and bar charts, such as Voronoi charts, tree charts, circular clusters, word clouds and so on. Although D3 can provide very fancy interactive charts, one thing to remember when choosing a data visualization tool is to know when to keep it simple.

7. Vision

If you need to make infographics instead of just data visualization, there are many tools available at present. Visual.ly is the most popular choice. Although the main positioning of Visual.ly is "the online market for infographic designers", it also provides a large number of infographic templates. Although there are still many limitations, Visual.ly is definitely a place that can inspire you.

Part III: Interactive Graphical User Interface (GUI) control.

What if data visualization is interactive enough to be used as a GUI interface? With the development of online data visualization, buttons, drop-down lists and sliders are all evolving into more complex interface elements, such as interactive graphic elements that can adjust the data range. When these graphic elements are pushed and pulled, the input parameters and output data will change synchronously. In this case, the graphic control and content have been integrated. The following tools can help you achieve these functions:

8. Cross filter

When we develop more complex tools to facilitate customers to browse data, we have been able to create small programs that are both charts and interactive graphical user interfaces. The JavaScript library Crossfilter is such a tool.

Crossfilter application: When you adjust the input range in one chart, the data of other related charts will also change.

9, tangled

The confusion of JavaScript libraries further blurs the line between content and control. In the application example shown in the figure below, Tangle generates an interaction equation of load, and readers can obtain corresponding data by adjusting the input value.

Part IV: Map tools

Map generation is one of the most difficult tasks on the network. The appearance of GoogleMaps has completely subverted people's understanding of online map functions in the past. MapsAPI released by Google allows all developers to embed map functions in their own websites.

In recent years, the online map market has matured a lot. If you need to embed a customized map scheme in a data visualization project, there are many choices in the market, but knowing when to choose which map scheme has become a key business decision. The map scheme looks powerful, but don't say, "With a hammer, everything looks like a nail."

10, schema mapping

As the name implies, ModestMaps is a very small map library, with a size of only 10KB, which is the smallest available map library at present. This seems to mean that ModestMaps only provides some basic map functions, but don't be fooled by this. With the cooperation of some extension libraries, such as Wax, ModestMaps will immediately become a powerful map tool.

1 1, leaflets

The CloudMade team brought another miniaturized map framework, Leaflet, to meet the needs of mobile web pages through miniaturization and lightweight. Leaflets and ModestMaps are both open source projects with strong community support, which are ideal choices for integrating map applications in websites.

12, multi-map

Polymaps is another map library, but it is mainly for data visualization users. Polymaps is unique in map stylization, and selectors similar to CSS style sheets are not to be missed.

13, open layer

OpenLayers is probably the most reliable one in all map libraries. Although the document annotation is not perfect and the learning curve is very steep, OpenLayers is unmatched for some specific tasks. For example, it can provide some special tools that are not available in other map libraries.

14, cartoon card

The marking line of the map is a rethinking of the map drawing. We are all used to Mercator projection, but Kartograph gives us more choices. If you don't need to call global data, but just generate a map of an area, then Kartogaph will make you stand out.

15、CartoDB

CartoDB is a website that can't be missed. You can easily associate table data with maps by using CartoDB, which is the best choice. For example, you can input a CSV communication address file, and CartDB can automatically convert the address string into latitude and longitude data and mark it on the map. At present, CartoDB supports the generation of five map data tables for free, and more users need to pay monthly.

With the popularity of high-definition mobile devices such as iPad3, the latest trend of web development is to integrate symbol fonts with fonts (change symbols into fonts) and create beautiful vectorized icons. Among these new fonts, such as FFChartwell and Chartjunk are specially used to display charts and graphs. Like OpenType, they are not supported by all browsers, but in the near future, these vector fonts will be considered in data visualization.

Part V: Advanced tools

If you plan to do some "serious" work on data visualization, then you may not be interested in online visualization tools or web applets. What you need is a desktop application and programming environment.

16, processing

Processing is a symbolic tool for data visualization. You just need to write some simple code and compile it into Java. At present, there is also a Processing.js project, which can make it easier for websites to use Processing without JavaApplets. Because the port supports Objective-C, you can also use processing on iOS. Although Processing is a desktop application, it can also run on almost all platforms. In addition, after several years of development, the processing community now has a large number of examples and codes.

17, node box

NodeBox is an application that creates two-dimensional graphics and visualization on OS X. You need to know Python programs. NodeBox is similar to Processing, but there is no interactive function of Processing.

Part VI: Expert Tools

Compared with Excel, it is a professional data analysis tool. If you are a professional data analyst, then you must know something about the tools that will be introduced below (if you are not proficient). As we all know, SPSS and SAS are standard tools in data analysis industry, but these tools are expensive and only large organizations and academic institutions have the opportunity to use them. Here we introduce several free alternative tools. The common feature of these open source tools is that they all have strong community support. The performance of open source analysis tools is not inferior to that of old professional tools, and the support for plug-ins is better.

18、R

As a statistical component package for analyzing large data sets, R is a very complex tool, which requires a long learning practice, and the learning curve is steepest among the tools introduced in this paper. But R has a strong community and component library, and it is still growing. When you can control R, all the efforts are worthwhile.

19、Weka

When you grow up to be a data scientist, you need to expand your personal ability from data visualization to data mining. Weka is an excellent tool, which can classify and cluster a large number of data according to attributes. Weka is not only a powerful data analysis tool, but also can generate some simple charts.

20、Gephi

Gephi is a tool for visual analysis of social graph data, which can not only process large-scale data sets and generate beautiful visual graphics, but also clean and classify data. Gephi is a very special software, and it is also very complicated. Mastering Gephi before others will make you ride the dust.