Use of Visual Analytics for Big Data

Abstract 

This study examined what big data means with the importance of it and the usage on each industry along with visual analytics to drive success in their organization. Various types of big data analytics tools such as Tableau, PowerBI, SAS etc. along with the comparison of the tools to discover best fit based on a profile of a company and it goals also are covered. We tried to examine how data visualization tools helped big technological giants to achieve competitive advantage taking care of the challenges that big data brings into visualization.

Keywords:  Big Data, Visual Analytics, SAS, Tableau, Visualization tools

Use of Visual Analytics for Big Data 

Have you ever wondered how much amount of data are we producing daily? The advancement of Web 2.0 and the growth of Internet of Things (IOT) is giving birth to around 2.5 quintillion bytes of data every day and the number seems to accelerate. Huge volume of data doesn’t equate to greater amount of information and to make sense of such huge amount of data the term “Big Data” was coined. It has been around since 2005 when O’Reilly launched it. Big Data refers to “datasets whose size is beyond the ability of typical database software tools to capture, store, manage, and analyze” (Manikya et.al, 2011). Its main characteristics is classified under 3 V’s which are Velocity, variety and volume. Combining big data with powerful visual analytics benefits an organization to achieve important business-related tasks. Companies utilizes big data to make smart decisions, cost and time reductions in the products and the methods used.

Industry-wise usefulness of Big Data

Personalization of content and recommendation of music analyzing the data based on previous choices made Spotify, an on-demand music providing platform, to drive business success. Not only the Media and Entertainment Industry but also other equivalent industries which were successful to utilize the collected data gained competitive advantage. Below, we will see how each industry benefits from the big data.

1. Banking: Bank uses big data to understand customers behavior and minimize fraud. Big data brings more insights of the data with more advanced visual analytics.
2. Government: This industry can be mostly benefited if strong big data analytics is used to map the differences of utilizing resources, managing crimes and handle national challenges such as terrorism, unemployment etc. with other minor and major issues of the government field with privacy and transparency in hand.
3. Health Care: Big data is helping this industry to get accurate results in terms of handling patients’ treatments along with speedy results. Also providing equal transparency to the doctor and patient record along with privacy can be targeted.
4. Education: Education industry might use big data to implement better system for evaluation of students and instructors.
5. Manufacturing and Retail: Introducing big data analytics in manufacturing industry will help to maximize quality while minimizing waste. And utilizing big data to retail industry helps retailers to learn the techniques to market customers and handle transactions effectively.
6. Media and Entertainment Industry: This industry can extract benefits from the big data for predicting the interest of its customers and target the effective advertisements based on personalization. Some of the tech giants which has incorporated big data re: Netflix, Starbucks, Amazon, T-Mobile, Capital One, American Express, Apple, Twitter, Google, Facebook etc.

In order to unlock the potential of collected big data it must be paired with some of the visual analytics to make proper insight of the data. “Visual Analytics is the science of analytical reasoning supported by interactive visual interfaces (Thomas, J, Cook, K, 2005)”. It makes insights easy to understand and take decisions. The primary importance of visual analytics is that: it can enhance decision making by learning about the past and implementing those learning in their new business stream.

Although Microsoft Excel which is popularly known for visualizing data has been there since ages, there are newbies which graphically represents data in a better way and by data, I mean BIG DATA. Similar task can be accomplished in multiple different ways, but depending on the size of the organization, volume of collected data and the performance outcome, we can choose from the list of available Visual Analytics tools: According to the survey published in International Journal of Computer Applications, the breakdown of the benefits of data visualization are shown in the table below:( M., S.  2011)

Now we have seen the industry usefulness of the visualization tools, let’s explore some of the popularly know visualization tools for big data which are:

1. Tableau: It is one of the mostly used data visualization tool in big data (Mahalakshmi R, 2015). According to the article published on Forbes, “Tableau has a huge customer base of 57000+ across industries” due to its simplicity and ability to connect to constantly changing dataset. It is highly reliable for huge changing datasets and visualizing big data and machine learning apps and can easily integrate with database solutions such as MySQL, Amazon AWS, Hadoop, SAP etc. Desktop, server and cloud-based online are the basic distributions available. Tableau is used by all data professionals but best for experienced users with knowledge on advanced machine learning.
2. Microsoft Power BI: PowerBI acts as an analytics platform for greater visualization capabilities as well allowing the option of creating visualization styles with developer tools. Power BI runs on Azure and connect to hundreds of data sources and simplify data prep. It is easy to use and mainly used for creating reports and dashboards than for complex tasks. 
3. SAS Visual Analytics: SAS uses predictive analytics to assess data driven decisions and provides interactive reporting, visual exploration and self-service analytics. It can be deployed on public cloud, data centers and on-premises. It supports advanced analytics allowing R, python, Java and SAS models.
4. QlikView: It is a powerful analytical and enterprise reporting tool that has good user interface and there is a great community to help new users.
5. Google Data Studio: A completely free Google solution to visualization is Google chart. It offers a wide range of visualization types and the tool has browser combability that absorbs real time data to create interactive dashboard.
6. Sisense: Sisense is yet another visual analytics solution that highly focuses on creating simple to complex interactive dashboards to give insight to the data for anyone working in the organization. AI and machine learning analysis is also integrated supporting large datasets.
7. Fusion charts:  This chart offers large number of templates with around 90 defaults chart toes and 900 maps to work with and can easily work with jQuery (I. B. Otjacques, 2013).
8. Plotly: Plotly would be the best choice for those who wants to work with more technical interactive visualizations along with the glitter of analytical programming languages like R, Python, MATLAB etc. Visualizations can be created directly or by using the libraries of the analytical languages. It is built on top of D3.js visualization library for JavaScript. The visualizations created can be easily embedded into websites or reports through HTML.
9. Carto: All of the visualization tools covered is suitable for small organizations as well as big but Carto its Software-as-a-Service (SaaS) model makes it easily affordable for small organizations and also scales accordingly. It doesn’t take long to map your data to the graphical interface.
10. JupyteR: It provides big data analysis and visualization across numerous programming languages. Jupyter Notebook also supports Java, Go, c#, Ruby in addition or Python and R. It easily interacts with Spark and process data from large data applications.

Every visualization tool has one thing in common that it provides insight to the data by improving response times, but the selection of the tools depends on the type of organization and its goals.

Techniques in Big Data Visualization

According to user requirements the visualization techniques are decided (Porter,2012). Conventional visualization makes use of tables, Venn diagram, entity relationship diagram, bar chart, pie chart for data visualization(Intel IT Center, 2013).Below is the list of visualization techniques for visualizing large amounts of data and getting insight about it are:

1. One-dimensional: It consists of one value per each data item or variable. Histogram is the perfect example of it.
2. Two-dimensional: As the name suggest, it has two variables. Bar charts, pie charts, scatter plot, maps are the type of 2D visualization.
3. Three Dimensional: This visualization will give more information to the user in the form of slicing techniques, Iso-surface, 3D bar charts etc.
4. Multi- Dimensional: It will give clearer picture of the visualization by analyzing the variables from different perspective. Parallel coordinates, Auto glaphs etc. are the type of such visualization.
5. Tree Map: Here the data is nested in form of rectangle which represents each branch of the tree. [3]
6. Temporal Technique:  It has the scalability of displaying the data in timeline, time series and scatter plot.
7. Network technique: It is used when you want to present data collected from the social media in the form of network.

FIGURE 1. Some of the visualization techniques

Challenges for Big Data Visual Analytics

The main challenge with visual analytics is to apply visual analytics to big data problems. Generally, technological challenge such as computation, algorithm, database and storage, rendering along with human perception such as visual representation, data summarization and abstraction are some of the common challenges.  “The top 5 challenges in extreme-scale visual analytics” as addressed in the publication by SAS analytics are as follows: (SAS Institute Inc, 2013)

1. Speed requirement: In-memory analysis and expanding memory should be utilized to address this challenge.
2. Data understanding: There must be proper tools and professionals who are proficient in understanding the data underneath the sea to make proper insight.
3. Information quality: One of the biggest challenges is to manage large amounts of data and maintaining quality of such data. The data needs to understood and presented in proper format that it increases the overall quality of it.
4. Meaningful output: Using the proper visualization technique according to the data presented is necessary to bring meaningful output to the data.
5. Managing outliers: While you cluster the data for favorable outcome, it is obvious that outlier will exist. Outliers cannot be neglected because it might reveal some valuable information and must be treated separately in separate charts.

Every data visual analytics tool that highlights its areas on theses majors will be able to gain competitive advantage amongst all the tools available. Recently, Hyper in-memory analysis was accomplished by Tableau. Now, we will see how some of the real benefits generated by visualization tools to help companies improve business and optimize performance.

1. Big companies like LinkedIn, Google, Apple, Southwest, Amazon uses tableau for their advantage. According to the article published in the Tableau website, “Tableau helped Southwest Airlines to maintain flights and optimize fleet performance”. It filled out the gaps and addressed the urgent request to ensure high fleet operation. Maiming fuel capacity and discovering the stations for tinkering helped Southwest to maximize cost savings
2. Another example is LinkedIn, Tableau helped LinkedIn to reduce customer churn. With Tableau, the increased product usage of a customer can be identified by a sales team. Tableau provided the best way to prevent the occurrence of churn.
3. Holding a share of 6.26% according to the report published on iDatalabs, QlikView was able to uncover trends to optimize operations across locations and deepen customer relationships for company ranging from Planet Hollywood to Lenovo to Deloitte etc.

Conclusion

As the number of generated data is growing rapidly on a daily basis, every organization will rely on data visualization tools that will help them to understand the trends and patterns to derive cost-effective ways to benefit their operations from those data. Many organizations are unable to choose the best visualization methods to undertake its operations which results in getting wrong outputs. Getting useful and right output will help the business to make perfect decisions and gain competitive advantage. This paper goes through analyzing big data and visualization tool basics, benefits of visual tools, challenges of those tools to implement big data with industry level outcome of those tools to gain success.

References

• J. Manyika, M. Chui, B. Brown, J. Bughin, R. Dobbs, C. Roxburgh, A.H. Byers, Big Data: The Next Frontier for Innovation, Competition, and Productivity, McKinsey Global Institute, 2011.Ren, L., Du, Y., Ma, S., Zhang, X. L., & Dai, G. Z. (2014). Visual analytics towards big data. Ruan Jian Xue Bao/Journal of Software, 25(9), 1909-1936. https://doi.org/10.13328/j.cnki.jos.004645
• Thomas, J., Cook, K.: Illuminating the Path: Research and Development Agenda for Visual Analytics. IEEE-Press (2005)
• Intel IT Center, Big Data Visualization: Turning Big Data Into Big Insights, White Paper, March 2013, pp.1- 14.
• M. Khan, S.S. Khan, Data and Information Visualization Methods and Interactive Mechanisms: A Survey, International Journal of Computer Applications, 34(1), 2011, pp. 1-14
• B. Porter, Visualizing Big Data in Drupal: Using Data Visualizations to Drive Knowledge Discovery, Report, University of Washington, October 2012, pp. 1-38.
• SAS Institute Inc., Five big data challenges and how to overcome them with visual analytics, Report, 2013, pp. 1-2
• Mahalakshmi R, Suseela S “Big-SoSA: Social Sentiment Analysis and Data Visualization on Big Data” International Journal of Advanced Research in Computer and Communication Engineering Vol. 4, Issue 4, April 2015.
• B. Otjacques, UniGR Workshop: Big Data- The challenge of visualizing big data, Report, Gabriel Lippmann, 2013, pp. 1-24.