Big data ai behind the scenes of Taobao’s Tmall Double 11: Real-time visualized query big screen

This picture comes from tmall Double 11 data live broadcast, from the charm of big data visualization

What is data visualization?

Taft says, “Graphs represent data. It’s actually more accurate and illuminating than traditional statistical analysis.” For the general editor, designer, operations analysts, big data, researchers, and so on all need from different dimensions, different levels and different granularity of data processing in statistics, with the aid of diagrams and information graph (for information only) for the user, the reader (consumption information) and managers (the use of information management and decision making) is different from the tabular analysis results. Data visualization technology makes comprehensive use of computer graphics, images, human-computer interaction, etc. It maps the data collected, cleaned, converted and processed in accordance with standards and specifications into recognizable graphics, images, animations and even videos, and allows users to interact and analyze data visualization. Any form of data visualization will be composed of three elements: rich content, eye-catching visual effects and fine production, which can be summarized as novel and interesting, substantial and efficient, aesthetic and pleasing to the eye.

【 Why 】 Data visualization?

Regardless of profession and application scenario, data visualization has a common purpose: to deliver information and knowledge accurately and efficiently, in a concise and comprehensive way. Visualization can transform invisible data phenomena into visible graphic symbols. It can establish connections and associations between complex and seemingly unexplainable data, discover rules and features, and gain insights and values of more commercial value. And use the appropriate chart to express directly and clearly and intuitively, to achieve the purpose of self-interpretation of data, let data speak. In humans, the right side of the brain can remember images a million times faster than the left side of the brain can remember abstract words. Therefore, data visualization can deepen and strengthen the audience’s understanding and memory of data.

Qun531629188 No matter whether you want to change your career or want to learn from college students, I am very welcome, today’s information has been upload to the group file, do not regularly share the trunk, There are also live streaming platforms for learning big data for free, including a new big data tutorial suitable for learning in 2018 compiled by myself. Welcome beginners and advanced friends.

【 How 】 How to achieve reliable data visualization

Data visualization including data collection, analysis, management, management, mining, a series of complex data processing, and then by the designer to design a form, may be 2 d graphics, 3 d view, regardless of the information, finally by the front-end engineers create the corresponding visualization and rendering and showcases the implementation of the algorithm. Just being able to turn data into pretty charts, and having fixed dimensions and different styles of charts to explain your point, doesn’t mean the end result is good enough. It was a simple beginning, a germ of good will. More work needs to be done to successfully report results and effectively turn the metrics and data you analyze into commercially valuable insights that can support fact-based decisions.

Color enhances the visual effect of information visualization. In information visualization, on the basis of clearly conveying information and narration through modeling elements, we should grasp the use of color in visual elements to make graphics more vivid and interesting, and to express information more accurately and intuitively. Color can help people classify information in depth, emphasize and dilute the expressive form of vivid and interesting visual works, and often bring visual effect enjoyment to the audience. Of course, the visual effect to the enterprise brand tone integration, and the enterprise brand culture to maintain a high degree of consistency, which is the most basic common sense. For example, if your company’s brand color is red, your visualizations should be consciously geared toward that tone. But there is no need to match, because the visual effect of red usually contains the charm of warning, so red is suitable for warning, reminding and highlighting information functions.

Typographic layout enhances the narrative of information visualization. I got wine. You got a story? Four basic principles of layout:

(1) Contrast: If two items are not exactly the same, they should be made different, and they should be completely different.

(2) Repetition: the Repetition of some aspect of a design throughout the work.

(3) Alignment: No element can be arbitrarily placed on the page. Each item should have some visual connection to something on the page.

(4) Closeness: It is used to organize related items together so that they are physically close to each other. The related items are considered as a cohesive group.

Dynamically increase the visual experience of information visualization. In the visual expression of information visualization, all kinds of information communication forms which are separated from each other are organically fused together dynamically for related and rhythmic information processing, transmission and realization. The ultimate goal is to explain the driving and related relationships between data representations in order to achieve linkage between data. Through the movement of chart style and color, the visual feeling of the audience is satisfied, and the information content is more profound and concise to the reader, so that the whole process of information transmission is easier and convenient. There are many tools for data visualization, such as ECharts, iCharts, D3js, Flot, Rapha? L and other functions are very powerful, but for non-professional visualization and often dealing with the chart of the workplace, a light and easy to learn and practical visualization software is very important. Like Cognos, Tebleue, etc. If you need to show the data structure is not particularly complex, and to show the data colorful, and interactive, then crystal table is the best choice.

1. Who are your readers?

Whether you’re making a traditional report or a new infographic, ask yourself first which readers are seeing the report? How much do they know about the matter to be discussed? What do they need? And how will they use the information and data you present? In “What is a Reliable Data analysis Report?” I talked about how important it is to have clear analysis goals and methods, because only clear analysis goals can have a good driving process. Whether it is goal-driven or process-driven, all the content presented in the subsequent data analysis work and analysis report are closely around the target theme and serve.

2. Plan a data visualization plan

Data visualization solutions must be able to solve user specific problems. Since it is able to solve user specific problems, then this height is based on your deep understanding of the phenomenon and nature of the data. Simply put, your visualization solution not only understands but is able to interpret the conclusions, information, and knowledge of the data analysis. And managers can quickly find and discover decisions along the visual path you’ve mapped out.

For example, when the performance of the enterprise is not up to standard (whether the performance of the enterprise is up to standard is related to the most critical interests and survival of the enterprise). The visual design path should look like this:

Step1, from the overall operation, identify the key factors that will affect the transaction and performance.

For example, the effective list, demo quality, customer service, product attributes, etc., and the performance of the KPIs corresponding to these key factors are the driving factors for the overall performance, and the KPIs corresponding to these factors have a direct driving and influence on STV. The visualization of these driving data is the foundation and the ultimate starting point and landing point for finding solutions. Because the performance of these data is the most direct view of the success of the operation.

Step2: make an in-depth analysis of key factors to determine what factors lead to the failure of performance, and find out the root causes and problems leading to the failure of performance.

Such as:

Through comparative analysis, the performance of KPI corresponding to all key factors from 201601 to December 2016 was observed one by one, and the difference of KPI corresponding to key factors in the month with the highest transaction performance and the month with the worst transaction performance was compared, so as to quickly locate the aspects and factors leading to the performance failure. It can then be targeted to drive and help the business to improve.

Track the implementation and progress of action plans that drive and improve turnover and performance, what problems exist, and whether poor implementation of action plans affects performance achievement.

Step3: in view of these factors, improve and explore ways to enhance performance with a definite aim.

Otherwise, there is no point in designing brilliant business visualizations without quick access to information and business decision recommendations and solutions. Visualization simply becomes the result of falsehood and deception, of pageantry rather than pragmatism. Based on the answers you have prepared for all of these questions, start customizing your data visualization solution to meet the specific requirements of each decision maker. Data visualizations should always be tailored to their audience, and such reports should include only what the audience needs to know, and put that information in a context that is relevant and meaningful to them.

3. Give your data visualization a clear title.

When your report is like a newspaper, magazine story. From this title, can give readers a strong impact. A clear title is a good explanation of the topic of the report and the story, and is a summary of the overall report and story. Of course, operations analysts are not encouraged to be clickbait. A good title is neither ambiguous nor superfluous, but simply explains the chart. This helps to get the audience straight into the subject. This allows the reader to scan the document and quickly grasp the core of the document. Try to make your headline stand out.

4. Connect data visualization to your strategy and plan

If the goal of data visualization is to introduce data that addresses specific, measurable, actionable, relevant, and time-sensitive issues, include those issues in your opening statement. This is later linked to your strategy to clarify the positioning of the data, so that the reader immediately understands the relevance and value of the visualized data. As a result, they are better engaged and able to use the information more wisely. Data visualization, in the end, serves for the good operation of the enterprise, which is its business value. It’s hard to build an infographic with linkage value if you don’t pay attention to your strategy and action plan. For example, the action plan implemented by the enterprise is usually to achieve and achieve the strategic goals of the enterprise, and lean management and operation can be achieved by such means. Therefore, the visual solution should be able to achieve the driving effect of action plan on strategic goals, and the driving and influence effect of individuals and teams on the overall indicators and KPIs of the department. Valuable data visualizations can only be achieved if a relational view of information is established.

Choose your presentation charts wisely.

No matter which type of chart is used, bar, line, radar, etc., each chart has its own advantages and limitations. You can’t find perfect visualizations. But you can make the visuals a little more human by trying to mix them up. All visualizations should convey information as simply and accurately as possible. This means that no matter how trendy, beautiful or gorgeous data visualizations are, they are not designed for that purpose. It is true that we are in constant and insatiable pursuit of the beauty of data. But the best balance lies in illustrating the value of the right information and knowledge with the right data visualizations.

Use only relevant graphics that convey an important message and that your audience wants.

Don’t fill up all the white space on the page — too much clutter interferes with important information, makes it too hard to remember, and too easy to ignore.

Appropriate use of color to increase the depth of information. Also note that some colors have hidden meanings. For example, red is considered the color of warning or danger. Suitable for early warning.

Don’t use too many different charts, tables and graphs. If you need to compare different charts, make sure you use similar charts to illustrate your data so you can compare them.

6. Add text instructions where appropriate

Text helps explain the data in words and adds depth to the content while contextualizing the chart. Numbers and tables may only provide a snapshot, but a text description gives more insight into the key points, comments on them and highlights them. Lead the viewer to think about the subject of the graphic, not about methodology, graphic design, graphic generation, or anything else.

Avoid distorting the original intent of the data.

Make a huge data set coherent.

Attract readers to compare and compare different data fragments, highlighting the key points and advantages and disadvantages.

The main idea should be fairly clear: describe, explore, chart, visualize self-interpretation.