Nowadays, big data is ubiquitous, and it is being applied more and more widely to history, politics, science, economy, business, and even penetrates into all aspects of our lives, and the channels for obtaining it are becoming more and more convenient.
Today we will talk about "big screen applications". When it comes to big screens, we must talk about data visualization. Nowadays, data visualization has become popular due to the popularity of data analysis, but data visualization is not a new technology. Visual data is data presented in a visual way. As a kind of big data display medium, the big data screen is widely used in various exhibition halls, exhibitions, press conferences and various carnivals, among which there are some common processing solutions: Ali's DataV, Baidu's Suger, Tencent RayData etc.
With the emergence and development of various connection-related technologies such as the Internet of Things and 5G, the amount of data held by everyone is increasing exponentially. It is impossible to understand these numbers alone. "Data visualization" is A simplification that turns the difficult data understanding process into - looking at colors, distinguishing lengths, and distinguishing high and low. This greatly reduces the time required to understand the data.
Because the company's self-developed products involve BI modules, the demand for large-screen data display was born (the demand for large-screen data has been completed).
The following are some exploratory practices I did in the early stage for this large data screen requirement, and the data is also mock
chart selection
The six basic charts cover most chart usage scenarios and are the most commonly used chart types for data visualization:
- Histograms are used to reflect the comparison between classified items;
- The pie chart is used to reflect the composition, that is, the proportion of the part to the whole;
- Line charts are used to reflect trends over time;
- Bar charts are used to reflect comparisons between categorical items;
- Scatter plots are used to reflect correlation or distribution relationships;
- Maps are used to reflect taxonomic comparisons between regions.
The basic chart types have common styles, but we will not explain too much. We think more about how to choose common charts to present data and achieve the goal of data visualization. Basic method: clarify the target —> select the graph —> sort out the dimensions —> highlight the key information .
Data request push
Once the information is ready, we need to fetch them from the server. Here we need a push-based approach such as the WebSocket protocol, polling, Server Push Events (SSE) and more recently HTTP2 Server Push. Here we briefly compare WebSocket and polling.
Polling requires the client to send an ajax request to the server regularly, and the server returns a response message after receiving the request. This requires a lot of server resources. At the same time, there are also some problems in the HTTP1.x protocol, such as head-of-line blocking and redundant headers. So this scheme passed directly.
Let’s talk about WebSocket, which is based on the TCP protocol. The data format is relatively light, the performance overhead is small, and the communication is efficient. It can send text or binary data. At the same time, it has no same-origin restriction, and the client can communicate with any server. Another point is that WebSocket usually does not use XMLHttpRequest, so there is no need to send header data every time we need to get more information from the server. This, in turn, reduces the costly overhead of data being sent to the server. This is undoubtedly the most efficient for large data screens that need to obtain data in real time.
Use more examples, and see other people's modules more