Data visualization in another buzzing word like data science and artificial intelligence in today’s world. In many a times, we are getting bombard with too many charts or graphs in many reports or presentation without proper information. On the other hand, few simple charts can tell us so many information or story behind it which we are experiencing number of times. So, it is important to choose right visualization for impressive and better impact. In this article, we’ll discuss that what is data visualization and how to create impactful visualization of data.
Table of Contents
What is data visualization
Data visualization is a method to represent the data in visual form with the help of charts, graphs, and diagrams. In other words, data visualization is the tool to create relationship among data, viewer, and designer. Let’s say, we have 20 years sales data of a product. Now it is very challenging to get insights by only looking into the numbers. But if we plot the 20 years sales data into a line chart, it is easy to find the insights out of it. Data visualization is so powerful by which data can be visualize and interpret well.
Let’s see the difference with an example. In the below graph, 20 years sales are plotted in a line chart. Insights can be drawn from the graph shown below are, sales were growing from the year 2000 to 2011. After that there was sudden dip in sales from 2013. During presenting the data, cause effect relation can be brought into it for better story telling. If only number are presented, it would be very difficult to interpret the data.
Characteristics of data visualization
- Data visualization should be visually impactful and appealing
- Data visualization should be saleable in nature
- Is should convey the right information to the viewers
- The dashboard or visualization should be easily accessible
History of data visualization
Data visualization is not a new concept, it is as old as humanity. We have seen many examples of data visualization in the pyramid or ancient documents. Ancient people tried to documented information in the form of visualization. There are many such examples of ancient data visualization like Papyrus Map of Egypt, Abraham Ortelius modern Atlas, map of cholera outbreak by John Snow and many more. In history, we have many such incredible work on data visualization which really added value in the field to visualize information.
There are few names who have contributed to the field of data visualization. Florence Nightingale created coxcomb or rose chart to represent the cholera data of mortality. Edward Tufte is called the father of data visualization. He was a statistician by profession and contributed to the field of data visualization.
Tools for data visualization
Data visualization tools are also called as Business Intelligence or BI tools. There are many data visualization tools are available in the market. These tools are helpful to create charts and graphs with the and create a dashboard and sequence of dashboard or storytelling.
Most Popular Data visualization or BI tools are Tableau, Power BI, Qlik Sense, Qlik View, Sisense, Looker, Google charts. All these tools are drag and drop tool and familiar with large data sets. Many other players are entering into the market with BI tools as the market has immense opportunity.
We should not forget here one basic data visualization tool that is Microsoft Excel. Excel has inbuilt and simple data visualization option to insert different type of charts and graphs. Data visualization functionalities are also available with R statistical tool and Python.
6 Effective Parameters to choose right data visualization
When creating data visualization from a data set, keep the following six 6 points in mind for an impressive aesthetic. These are considered as standard aesthetics for data visualization.
- Position
- Shape
- Size
- Color
- Line width
- Line type
All these 6 parameters will be discussed in four part, choosing right color, choosing right font, choosing right graph and additional tips for data visualization.
Choosing the right color
Choosing right color is an important parameter for data visualization or creating a dashboard. Many a times, people are using too many colors in a dashboard which dilute the story or purpose of the dashboard. So, color is an important component of data visualization and choosing right color means right impact.
There are three types of color segment in data visualization:
Categorical color: Categorical color is different color from each other, and it is to show comparison mainly for categorical variable.
Diverging color: Diverging color is having three color, two polar color and one middle color. If having large variance in data, diverging color needs to be used. If plotting sales achievement ranges from 0 to 100, 100 and 0 should be tagged with polar color and 50 should be with middle color.
Sequential color: Sequential color is range is having one main color which is dividing into two color one is deep and another is light. As an example, light blue and deep blue is the example of sequential color as shown below. Sequential color is used in data visualization where data range is showing from low to high.
Apart from these three types, another type is called highlight color. Highlight color is to highlight a specific portion of data to stand out or creating an alert.
Keep color blind factor in mind:
When you are preparing a dashboard with the help of data visualization, most of the times it is for a group of people or team members. Now think of a situation where you have used red and green color in a scatter plot to denote different category of product (fig 1.2). In the group, 50% people have CVD (Color vision deficiency), that means those are not able to differentiate between red and green. Your presentation will not have any value for them right! For people with CVD, both red and green color will look like brown.
So, as a good presenter, you need to convey your insights from data to everyone and that’s why it is important to keep the CVD factor in mind. As per a research done in 1993, 8 percent of the male population have CVD. On the other side, 0.4 percent of female population have the same deficiency.
There are three types of CVD, Protanopia (poor Red color vision), Deuteranopia (poor green color vision), Tritanopia (poor Blue color vision).
It is not only red and green because it is associated with color frequency. As a result, orange color is also look like brown for people with CVD.
As a solution, don’t use red, green, orange, and brown color simultaneously in a graph. It is better to use any one color from these and other colors.
Choosing the right font
Data visualization is not only creating and choosing right graph and color but also choose right font of text and numbers.
Find the following points as good practice to choose right font.
- Use maximum 2 types of font.
- Roboto, Lato, Noto and Sans are preferred choice of font for data visualization or dashboard. For headers, PT Sans and Lora are preferred choice. It is for good visibility by the readers.
- For headings, increase the font size a bit and make it bold for differentiation.
- Deep and light color text can be used to differentiating and make visual impact.
- Keep good letter spacing and distance between two lines so that the text can be easily readable.
Choose the right graph or visualization
There are many types of graphs available in data visualization. Type of graph has been chosen based on the objective or what do you like to communicate to the audience. If right graph is not selected, it may not be communicating the right message to the audience. Hence, following cases guide you to understand which graphical representation is appropriate for which type of storytelling.
Show Comparison: To show the comparison among data or variables, four types of chart can be used.
- Column chart
- Bar chart
- Line chart
- Scatter plot
Show change over time: To show the time series data or data trend over time of quantitative data, following charts should be used.
- Line chart
- Bar chart
- Area chart
- Column chart
Distribution of data: To show the distribution of data set, following charts should be used. Distribution type of charts are also called statistical chart.
- Histogram
- Bell curve
- Boxplot
Visualize Correlation: To visualize correlation between two variables, below charts should be used.
- Scatterplot
- Bubble chart
Composition or breakup: To show the composition or breakup, following charts should be used to visualize.
- Pie chart
- Donut chart
- Cox-comb chart
- Rader chart
Project Management: In project management, most useful chart is Gantt chart to show the progress with start and end date of the project. Gantt chart is so important for project managers that’s why it is part of think-cell tool.
Show Rank: To show the rank in a data set Bar chart is the best way to visualize.
Show Hierarchy: Tree map is being used to show the hierarchy exist in a data.
Geography chart: Map is most effective way to show the geographical data instead of only name of geography. In this case, data is being plotted on a map in the form of size, shape and color.
Relation among data set: Venn diagram is the method to show relationship among two or more data set. To explain join concept in SQL or relational database concept, Venn diagram is used.
Show Process: Waterfall chart is the best way to show step by step process.
Show Intensity: In case of large data, it is difficult to visualize the intensity of quantitative data. So, heat map is being used to show intensity in a data.
Additional tips for data visualization
There are terminologies associated with data visualization which is associated with the concept and making data visualization more meaningful, easy to understand and good aesthetics.
Data Source: Always it is good practice to mention data source or data visualization created by using which data.
Axis Name: Mostly 2-dimensional graph is used for data visualization. So, it is mandatory to mention the axis name for X axis and Y axis.
Scale of Graph: One characteristic of data visualization. So, it is important to use right scale in a graph. Mostly scale interval for both X axis and Y axis is presented with the interval of 10.
Legend: Legend add more clarity in case of complex graph where multiple variables are accommodating in single graph. Multiple variables are being showed in a graph by using different color, so which color denote which variable that needs to be clear by using legend.
Graph Size: Size or area of a graph is crucial for visualization. If it is too small, not visible clearly. On the other hand, if it is too big, not looks good. So, it is important to keep a good graph size based on the content, distance of audience and number of details in a graph.
Line type and size: Line is used in data visualization as separator or border. It is better to use small and dotted lines to create a nice looking visualization.
Conclusion
In this article, we have discussed all the required details of data visualization. To master in the topic of data visualization, see as many as data visualization template or dashboard. Also, create your own dashboard or template by using different data sets to practice. As many as you practice and explore other’s work, you will become master of this skill.
All the best.
