1. Explore variable correlations
Determine whether two variables change in sync, providing a basis for subsequent modeling (such as regression analysis).
2. Detect data distribution patterns
Identify linear, exponential, logarithmic, and other relationship types. For example, population growth and time may show an exponential distribution, while car speed and fuel consumption may show a U-shaped relationship.
3. Identify outliers and abnormal values
Outliers may affect model accuracy and require further investigation (such as data entry errors or special events).
4. Compare differences between multiple groups of data
Use group labels (such as different colors) to compare the differences in variable relationships between different groups (such as different age groups).

1. Observe the overall distribution pattern
Linear trend: Points are distributed diagonally upward or downward, such as height and weight.
Nonlinear trend: Points are distributed in a curved line, such as car speed and fuel consumption.
No correlation: Points are randomly scattered, such as height and shoe size (after controlling for age).
2. Identify outliers and abnormal values
Examine points that deviate from the overall distribution to determine whether they are data errors or special cases. For example, when analyzing the relationship between housing prices and area, a location with high housing prices in a small area may be in the city center.

1. Analyze variable correlations, such as whether advertising investment and sales increase or decrease in tandem, to quickly determine positive/negative correlation or no correlation.

2. Detect data distribution patterns, such as identifying nonlinear relationships (such as a U-shaped curve between car speed and fuel consumption) or outliers (such as a small, high-priced outlier).

3. Conduct multi-group comparative analysis, distinguishing groups by color/shape (such as differences in height-weight distribution between men and women), to assist in market segmentation research.

4. Validate regression models, checking whether variables meet linear assumptions and avoiding model bias.

The key difference between scatter plots and bubble charts lies in the variable display dimensions and visual expression. Scatter plots use a two-dimensional coordinate system to display the relationship between two variables (X, Y). Each point represents only a pair of values. They focus on analyzing correlations, trends, or outliers between variables and are suitable for exploring basic relationships. Bubble charts, on the other hand, introduce a third variable, encoding values ​​through the size (or color) of the bubble, enabling simultaneous analysis of three variables.