Top 6 simple & efficient visualization types briefly

If you want to discover or present your data, you came to the right place

Margarita Arutiunova
CodeX
4 min readApr 4, 2022

--

Photo by Carlos Muza on Unsplash

We have several ways to understand data. When we’re analyzing it, we think about visualization last. Nevertheless, our minds are set in the way we need the visual form of things we want to examine. Therefore, visualization is essential not only for presenting some conclusions but also for the world patterns discover.

Even working with some numeric information not related to everyday-life stuff, we often need to find some sequences and patterns in the data to analyze it. If we see the picture, we can do it faster. Thus, the fundamental purpose of visualizations is to create a visual form for a better and more efficient understanding of patterns hidden in the data.

As a bonus: the visualization can illustrate the written reports or articles for providing some ideas to readers easier.

Nevertheless, the article is devoted to the top of simple visualization types. Therefore, it’s my pleasure to share the brief collection of visualization options I use almost every day:

  1. Box Plot
Image by author

This tool shows the basic statistics. It is effective for observing a single variable. We can split it into multiple box plots by a categorical variable. In this case, we can compute ANOVAs or chi-squares and order the variables by how well they split the categories.

2. Curve

Image by author

It visualizes distributions, but it should be considered carefully as this tool is some smoothing result. It, accordingly, has some parameters that are hidden. Sometimes, they are hard to set and/or interpret. Sometimes, curves result from parameters fitting to a shape, which we have to decide what should look like. We could have no good argument for a particular shape vision in this case.

3. Histogram

Image by author

This tool is also effective for distributions. Here, we also can split the data by a categorical variable.

Note: the most significant issue that can be here is choosing the number of bins. This feature can strongly affect the visualization’s final form.

However, histograms are better than curves essentially because they do not hide anything. If not the criteria for the number of bins or boundaries decided are manipulated, it works efficiently.

4. Mosaic Display

Image by author

This tool is useful for relations between categorical features in a similar fashion discovering as the scatter plot does for numeric variables. It splits the data by 1–4 variables into smaller subgroups. Mosaic Display performs the area size corresponding to the group size. This area can be split further to show the target variables’ distribution inside the group.

5. Scatter Plot

Image by author

As it was said in the previous point, Scatter Plot shows the relationship between two numeric variables. Here we can use sizes, colors, shapes, or even angles to show other quantities in the same plot. The colors are the attribute that works more effectively out of the first three mentioned. It is easy to see the concrete regions with a color’s prevalence in most cases. It also allows spotting patterns where are none there.

Note: using colors is the most effective way to show numeric quantities. The main point is to use a discrete scale with not too many colors. It would be hard to compare a color in a legend with the spot’s one on the graph if the colors scale is continuous.

6. Sieve Diagram

Image by author

This tool shows the two variables’ actual frequencies against their expected frequencies. It works as a chi-square visualization using colors and a grid. The area is split independently along each axis, unlike the Mosaic display. Therefore, assuming the variables are independent, the sizes correspond to the expected instances number. This type of diagram also shows the independence violations. In other words, it opens us to the combinations of the values that are more rare or common than expected.

Thus, I tried to perform the top 6 simple & efficient visualization types as briefly as possible. Each of them is suitable for different cases. The crucial point here is the needs and purpose of data analysis determination. According to that, choose the most appropriate visualization type.

Thank you for reading! If you want to share your opinion or ideas with me, you can write freely in the comment section. Feel free to reach me on LinkedIn profile for any suggestions or clarifications.

Have a nice day!

--

--

Margarita Arutiunova
CodeX
Writer for

Margarita Arutiunova has experience in Marketing, Machine Learning & Data Science. She holds a Master’s degree in Managemant and Analytics for business.