Smart Analytics | Blog | Our Top 9 Ways to Visualize Your Data

June 26, 2020

Our Top 9 Ways to Visualize Your Data

Here we will talk about data visualization. We will show you a selection of analytical reports from our solutions where various data visualization tools have been used - from basic to advanced ones.

What Is Data Visualization for?

People perceive and remember information better when it is visual. It looks more usual and understandable when represented as attractive charts, graphs and diagrams.

Data are visualized to represent the nature of a concept or the condition of a process under our scrutiny. Data visualization allows us to get the picture of the situation here and now without losing ourselves in figures and tables.

With raw data converted into a graphical format, it is easier to explain complex issues to deal with business tasks like the analysis of the performance in specific actions, the analysis of task dynamics, the analysis of reaching objectives. The success of visualization hinges on its adequate use, namely the correct choice and reasonable use of graphs and diagrams.

You are free to choose which method of data interpretation will suit you best. First we will go over the traditional ways of data visualization on the example of the solution UN ESCWA Data Portal, and will then move on to more advanced ones on the example of the solution 3M™ Benchmark Portal.

The Most Popular Types of Data Visualization

You can see above several examples of the traditional ways of data visualization taken from the Open Data Portal developed by Smart Analytics for the UN ESCWA. The users of the portal are any visitors interested in studying social and economic statistical data from 18 countries of the Arab region.

Pic. 1. Pie Chart

A pie chart is one of the most popular and easiest ways of representing figures graphically. Such diagrams are easy to read and understand because it is clearly visible how parts relate to the whole. Pie charts are ideal for promptly letting you grasp the idea of the proportions of the data segments. In our example you can see the division of CO2 emissions by industrial sector.

Pic. 2. Bar Chart

A bar chart is another popular way of visualizing data that easily gets the message across. In a bar chart a dataset is transformed into bars whose height depends on the proportions of the represented values. These diagrams facilitate sorting data in one category within a limited amount of time. It could be used for demonstrating the changes in the number of the population of a country within a given period. In our example it is the dynamics of structure changes in the GDP of South-West Asia in the past 18 years.

Pic. 3. Line Chart

Just like bar charts, line charts help visualize data accurately in a compact way, which allows to easily get the message across, define trends and collate indicators (when using several lines). Line charts are used to display resulting data relative to a continuous variable, usually a time series. This type of visualization calls for adequate use of colours to make information analysis still easier for the user.  In our example we have a line chart showing the number of passengers travelling to or from countries of the Arab World in the period from 2005 to 2018.

The Data Portal allows combining tables, charts, graphs and texts.  A tool like this for an attractive and user-friendly dissemination of data never existed before at ESCWA.

Statistics Division

UN ESCWA

Data Visualization Types for More Detailed Analysis

Sometimes simple data visualization does not yield enough detail thus not quite fulfilling its task. In order to specify information and display crucial details beside the graph or diagram advanced ways of data visualization are used.

In the given example we show the solution developed for 3M Health Information Systems. It is Benchmark Portal which provides users with efficient and intuitive tools to make decisions in managing costs, analysing the performance of Belgian hospitals and benchmarking these hospitals. The solution can be implemented in other fields as well.

Pic. 4. Sankey Diagram

A Sankey diagram is a specific type of visualization used to display the flow of one set of values to another. This kind of diagram is often used to demonstrate business processes or dataflow. A Sankey diagram shows workload, capacity, efficiency, as well as correlations, their power and contribution to the common flow. The lines in the diagram can overlap or diverge. The width of a line depends on the ratio in the flow, with the line thickening as the parametre increases. The diagram is also often used to demonstrate the ratio of budget costs and budget revenue. In our example the diagram shows two models of hospital financing: the old model is on the left, the new variant is on the right. It helps analyse the changes in the structure of financing after assuming the new model’s set of rules.

Pic. 5. Radar Chart

A radar chart, also known as a spider diagram, comes in handy for collating quantitative variables. It is possible to easily compare several indicators at the same time and see if there are any serious digressions that may require further research. In our example the radar chart demonstrates healthcare care profiles in a hospital. It shows the frequency of treatments in one and the same department – e.g. in intensive care. The higher the frequency of treatments, the more costs would need to be allocated for this or that treatment.

Pic. 6. Spider Chart

Another analytical report with the use of a spider diagram can be seen in the next example. This example demonstrates the generalized situation at the hospital at the key points of the quality of treatment: repeated admissions of patients, medical complications, mortality rate and average duration of stay in hospital. The red line shows the average indicators, the dark-blue line shows the current situation and the dotted line demonstrates the situation in the last quarter.

Pic. 7. Scatter Plot

A scatter plot, or a scattering diagram, is one of the seven key graphical methods most suitable for handling quality issues. It looks like a graph with values marked on the X axis and Z axis where every pair of values can be represented by a dot. The accuracy of the model may be judged by how scattered the dots are on a given graph. If most dots accumulate along a line and digressions are negligible, the model works well. On the other hand, the accuracy of the model is not so good if the dots are widely scattered.

Every line is a separate scatter diagram in the screenshot. This analytical report reveals the level of coding errors in certain categories of treatment in a hospital, with every category involving treatment for a certain group of medical conditions. The report also shows the quality of coding in hospitals. The red circles are average values while the blue ones show the situation in this or that hospital.

Pic. 8. Heatmap

A heat map serves to show a large volume of comparison data where values are displayed as marked coloured tiles. Heat maps are convenient when it comes to showing variations with a number of variables, detecting patterns, revealing similarities between variables and uncovering any dependencies among them.

A heat map is ideal for comparing the efficiency of a company’s divisions, as well as for detecting priorities for investments or areas for improvement. In our example the map shows the prevalence of patients’ diagnoses for groups of medical conditions.

Pic. 9. Violin Plot

The last type is a violin plot. This statistical diagram is used to visualize the division of data and their frequency distribution. Every ‘violin’ represents a group or a variable. Its shape shows the estimated frequency distribution for a variable. The more dots there are in a given range, the more violin expands. Even though the graph may appear peculiar, it often proves to be quite useful, especially when the shape of distribution is called for.

A violin plot is used to show frequency distribution comparison in different categories. It helps the user to decide whether an average value can be used instead of a probability estimate.

In our example every violin is symmetrical from left to right. Although ideally these violins are supposed to be compact with very short tails, in our case the violins are elongated with the median value rather stretched out. This is a signal that the choice of the categories was not ideal.

We heard nothing but good things about our portal.

Users of the Portal

3M Health Information Systems

Data Visualization for Tasks of Any Complexity

Data visualization is a powerful tool for demonstrating the results of business analysis that helps to clearly show to users data and gives recommendations.

With a wide range of options of how to visualize information – from pie charts and bar charts to radar and violin plots – data become clear and visually compelling for the user. The correct choice of visuals for the analytical report may reveal better than anything what the data really contains. It should be taken into account that the data should be clear to all categories of users – analysts, managers and any interested external users.

Solutions for data visualization and dissemination developed by Smart Analytics facilitate working with a wide array of data sources. They allow users of any level, from beginners to advanced data analysts, to analyse data on their own, as well as build interactive reports of any complexity and share information with interested users worldwide.

Tags

  • Data Visualization
  • Dashboards
  • Data Analytics
  • Business Intelligence

Subscribe to the Blog

We will share with you our knowledge and insights. Spam-free, only useful content.