12 types of practical charts that will be used for data analysis
"The auxiliary line drawing is so fancy, it's useless if you don't get the title."
The same goes for the workplace.
Beautiful ≠ good. The essence of data visualization is to effectively convey information, so for data analysis, the practicality of charts always comes first.
1. Basic chart
Pie chart, histogram, indicator chart, bar chart, line chart, detailed table.
Different types of charts have different presentation methods, and the key content expressed is different.
Generally speaking, data usually contains five kinds of correlations: composition, comparison, trend, distribution and connection.
- Composition -Mainly focus on the percentage of each part of the whole. If the information you want to express includes: "share", "percentage" and "what percentage is expected to reach", you can use [pie chart] at this time
- Comparison -can you show whether the order of things is similar, or is one more or less than the other? "Greater than", "less than" or "approximately equal" are keywords in the comparative relationship. At this time, the first choice will be [bar graph]
- Trend -care about how the data changes over time. The weekly, monthly, and annual trends are increasing, decreasing, fluctuating up and down or basically unchanged. At this time, use the [line chart] to better display the indicators over time the trend of
- Distribution -It is concerned with how many items are included in each numerical range. Typical information will include: "concentration", "frequency" and "distribution", etc. At this time, use [Histogram]
- Connection -mainly to check whether the two variables express the model relationship we expect to prove. For example, the expected sales may increase with the increase of the discount rate. At this time, you can use the [line chart] to display it to express " The relationship between variables related to, "growth with ...", and "different with ...".
2. Derivative form
Because of the different application scenarios, more forms are derived from the basic diagrams to give richer expressions to different data relationships. For example:
from a histogram to a stacked chart, it is not difficult to find that the stacked chart shows a share relationship based on the histogram.
The area chart emphasizes the extent to which the quantity changes over time, drawing attention to the trend of the total value. Area charts can also be "stacked" to increase the concept of share.
Observing the change process from line chart to scatter chart, from scatter chart to bubble chart, we can find that the scatter chart has more space to express the dimension of the data. As a result, the range of "scattered points" in the bubble chart is variable.
3. More concrete charts in other scenarios
Use the map as the base to show the dynamic distribution of the area:
Gantt chart is a very commonly used chart type in project/task management.
As shown in the figure, the order and duration of specific projects are shown through the activity list and time scale. It is the leader in project management.
Horizontal comparison, showing insufficient radar chart:
Four, chart matching
All kinds of charts are powerful, and they are more effective through reasonable matching:
Bar chart + line chart
Show data association in two forms.
Thin ring pie chart + indicator chart
While the performance accounted for, amplify the important indicators.
Map + bubble chart
Show the severity of the epidemic situation in various regions of the world.
In summary, the combination of indicator charts, histograms, bar charts, radar charts, etc. can display data analysis results in multiple dimensions.
At this point, the chart becomes clearer and more interesting.
However, I still want to remind everyone not to pursue beautification excessively, but to practice more. The best way to understand chart knowledge is to see more, use more, and practice more.
Some of the pictures of the epidemic in the article came from the Internet and were deleted.
Tools used in making charts in the text: Jian Daoyun