Overview:
The enormously-grown data and capability to analyze this humongous data demanded visualizations or visual reports that are more intuitive and interactive than the traditional reports. There are many BI reporting tools came up with the new and effective styles of visualization including the advanced geospatial visualization. However, what kind of visualizations are best suited to present a dataset? Does one style depict all business metrics effectively? Traditionally, Pivot tables, Histograms and Pie Charts were among the other few graphs used by almost all the businesses and different departments in a business. In the last few years, The businesses found other visualizations like heat maps, tree maps and network visualizations more comprehensible.
In this post, let us randomly (not really) choose 3 business vignettes and discuss about the best ways to visualize metrics in those businesses.
Human Resources Management:
Human Resource Management, recently evolved as Human Capital Management with many additional features, will be the first business vignette we look at. Before exploring the best visualization, let us examine what are the metrics or data the reports consist of in this domain. One of the most common reports in this realm will be Employee Turnover Report (staffing analysis) along with Employee Transactional Reports, Employee Skills, Employee demographics including age and Time & Labor reports.
When it comes to the Employee Turnover Report, I believe the best way to look at it will be a line graph depicting the staffing activities by Department a period of time. Tools like Tableau® has an option to create beautiful dashboards including staffing trends, performance by groups etc.
source: http://www.tableau.com/solutions/hr-analytics
In this example, the line graphs display the number of hires/rehires, promotions, terminations and transfers differentiated by color making it easy to pursue. The stacked histograms below explains the performance details by supervisor, percentage of employees’ performances and terminations in each section of performances respectively.
An extended version of this report will be an ‘employment report’. Looking at a time-series plot of the employment trends is a quick and better way to grasp the employment trends over a period of time, generally published by expert firms like ADP LLC and Forbes.
source: http://www.adpemploymentreport.com/2015/February/NER/NER-February-2015.aspx
Healthcare:
Healthcare is one of the complex systems in United States that needs informative visualizations for a variety of users including PCPs, Hospitals, Insurance Carriers and government. The metrics in this domain include monetary details like average expenses on healthcare, premiums or hospital bill amounts along with other measures like percent of people insured. Further analysis based on the geographical locations make healthcare analytics more interesting.
Though all the numerical values can be shown in histograms effectively, I prefer the bubble plots which are aesthetically appealing to eyes.
Here is an example of aggregated healthcare expenses shown in bubble plot.
source: http://www.knowledgevaluechain.com/wp-content/uploads/2013/07/Bubble-chart.png
Another chart, I found very interesting is relating the analysis to geographical locations and presenting in a geospatial heat map like below.
source: http://www.pewtrusts.org/en/multimedia/data-visualizations/2014/medicaid-spending-growth
E-Commerce:
E-commerce is another business which has adapted the new Big Data and Data Analytics technologies at their nascent stages. This industry has many quantitative and qualitative metrics to be visualized in the plots. The quantitative metrics include session time, conversion rate, average order value, churn rate, cost per clicks (CPC) and customer lifetime value (CLV) and not to forget the traditional financial and revenue numbers.
I see histogram, line graph or scatter plot as feasible and good visualizations for presenting quantities like conversion rate over time grouped by day and time. However, this heat map made it simple and more attractive.
source: https://plot.ly/~hianalytics/16/google-analytics-organic-traffic-conversion-rate-heatmap-jan-2014.png
All the other rates and numerical parameters can be presented in a similar plot. Adding to the list is a word cloud which presents the keywords that attract more customers to the website. Here is an example:
source: http://searchenginewatch.com/IMG/854/245854/enterprise-search-2013-word-cloud.jpg?1358369836
Final word:
The views expressed in this blog are purely my personal opinions and I am sure there are many other interesting and attractive plots to display the same or similar metrics.
Wear your creativity hat and generate amazing visualizations!!
References:
The Data Warehouse Lifecycle Toolkit - Ross & Kimball
http://www.adpemploymentreport.com/2015/February/NER/NER-February-2015.aspx
http://www.tableau.com/solutions/hr-analytics
http://www.aquire.com/software/workforce_analytics_old
http://www.knowledgevaluechain.com/wp-content/uploads/2013/07/Bubble-chart.png
http://www.pewtrusts.org/en/multimedia/data-visualizations/2014/medicaid-spending-growth
http://blog.bigcommerce.com/6-vital-ecommerce-metrics/
https://plot.ly/~hianalytics/16/google-analytics-organic-traffic-conversion-rate-heatmap-jan-2014.png
http://searchenginewatch.com/IMG/854/245854/enterprise-search-2013-word-cloud.jpg?1358369836
