Data Visualization Framework

At present, big data and machine learning are extremely hot. Every day, we can see news that various AIs replace certain human occupations. Data is a carrier that contains a certain combination of information. In the final analysis, big data is to extract a large amount of data to form a natural language that humans understand and provide support for people to make business decisions.

Images are the best carrier for humans to understand things, so how to convert icy data into beautiful charts becomes very important. Good charts can achieve half the result with half the effort in displaying products or internal PPT responses.

The data of the category comparison chart generally contains two types of data: numeric and category. For example, in a column chart, the X axis is category data, and the Y axis is numeric data, which uses two visual elements: position + length. Category data mainly includes column charts, bar charts, radar charts, slope charts, word clouds, etc., which are usually used to compare the scale of data. It may be a relative size (which shows which one is larger), or it may be an absolute size (need to show the exact difference).

Category comparison
Data relationship charts include two main categories of charts showing data correlation (Correlation) and data flow (Flow).

Data flow charts mainly show readers the flow or intensity of two or more states and situations, including network charts, chord charts, sangki charts, honeycomb charts, etc. Among them, the network diagram is to show the relationship strength and internal relationship between different types of objects.

Data correlation charts mainly show the relationship between two or more variables, including the most common scatter charts, bubble charts, surface charts, and matrix scatter charts. The variables of the chart are generally numeric. When the variables are 1 to 3, scatter charts, bubble charts, surface charts, etc. can be used; when there are more than 3 variables, high-dimensional data visualization methods can be used, such as parallel Coordinate systems, matrix scatter plots, radial coordinate plots, star charts, and Cherzhov masks, etc.



Data relationship
The data distribution chart mainly displays the numerical values in the data set and their occurrence frequency or distribution law, including statistical histogram, nuclear density curve, box diagram, violin diagram, etc. Among them, the statistical histogram is the simplest and most common, also known as the mass distribution map. A series of vertical stripes or line segments with different heights represent the data distribution. Generally, the horizontal axis represents the data type, and the vertical axis represents the distribution.



Data distribution


Time series charts emphasize the changing rules or trends of data with time. The X axis is generally time series data, and the Y axis is numeric data, including line charts, area charts, radar charts, daily force charts, and bar charts. Among them, the line chart is a standard way to display the trend of time series changes, which is very suitable for displaying the trend of data at equal time intervals.



sequentially
The partial overall chart can display the information of the local composition and the overall ratio, mainly including pie charts, doughnut charts, sun charts, waffle charts, and tree charts. A pie chart is a common way to show the relationship between parts and the whole. In a pie chart, the arc length (and center angle and area) of each sector is the ratio of the number represented. But it should be noted that it is difficult to accurately compare the size of different components in this type of diagram.



Partial whole
Geospatial charts mainly show the precise location and geographical distribution of data, including maps of equivalent intervals, maps with bubbles, and maps with scattered points. The geographic coordinate system for maps can map location data. There are many forms of location data, including longitude, latitude, and zip code. But it is usually described by latitude and longitude.