Linking Big Data Visualization to the Value of KPIs
Given that we live in a time of data and information overload, we really need to have better mechanisms to make sense of it all. Data visualization supports the transfer of information to knowledge by illustrating hidden issues and opportunities in big data sets.
Big data is creating unrivaled opportunities for businesses, aiding them to achieve faster and deeper insights that can strengthen the decision making process, improve customer experience, accelerate the speed of innovation and gain a competitive advantage.
A significant amount of the human brain is dedicated to visual processing, resulting in our sight having a sharpness of perception far surpassing our other senses. Effective data visualization shifts the balance between perception and cognition. The visual cortex is used at a higher capacity and the viewer understands the presented information much quicker, better and is able to make a superior decision based on the findings.
Businesses are increasingly turning to data visualization to discover the overwhelming amount and variety of data cascading into their operations, and to eliminate the struggle of just storing the data and focus on how to analyze, interpret and present it in a meaningful way. The trend towards data visualization is worth delving into by any business seeking to derive more value from big data.
The IBM framework for tackling big data focuses on “the four “V’s” (volume, velocity, variety and veracity), but does not emphasize enough another “V” that requires attention, namely visualization. Even with the use of business intelligence tools and the incredible exponential increase in computer power, the need to consume information in a meaningful manner exceeds the ability to process it.
Visualization plays a key role starting from the raw input of the big data, where structures and underlying patterns that may be held within the data can be observed or formed, to the end result of a visual representation that presents valuable key insights in a fast, efficient and clever manner.
Crafting a visualization is more than simply translating a table of data in a visual display. Data visualization ought to communicate information in the most effective way, with the prime purpose of truly revealing data in a quick, accurate, powerful and long-lasting manner.
The main problem with big data involves complexity. Information and data is growing exponentially with time, as an increasing amount of data is made available on the internet. Furthermore, the number of insights, opportunities or hypotheses hidden in a dataset is exponential to the size of the datasets.
In achieving efficiency and ensuring the comprehensibility of visual representation resulted from big data, KPIs can be used, as to attain the goal of graphical excellence and to add value to the end result. Big data visualization requires skills that are not intuitive and the entire process relies on principles that must be learned. Each big data visualization created should follow a clear path to success, namely: attain, define, structure, extract, load, display, refine the data and interact with it.
Key performance indicators add value to the entire process by ensuring clarity in developing the strategy of the project, focus on what matters and requires attention, as well as improvement by monitoring the progress towards the desired state.
When developing a project of big data visualization, the process should follow a cursive and pre-defined flow in accordance with the project needs and the requirements of the end-user. These recommended stages are:
1.Acquiring the data: this is usually how the process starts, unrelated with the platform which provides the data. In the process of big data collection, there is also the issue of data selection. Instead of “just throwing it all in”, one should focus on selecting high quality data, which is relevant to the project’s objective and does not add noise to the end result.
The noisier the data is, the more problematic it will be to see the important trends. It is suggested to have a clear strategy for the data sources required, as well as for the subsets of data relevant to the questions the project wants to answer to.
2.The next phase in the process refers to structuring of the acquired data. This includes the process of organizing the data to align it to one standard. The data store might be comprised of a combination of structured, semi-structured and unstructured data.
At this stage, it is easier to identify the common aspects in each sets of data and to find relationships between the data at hand. This includes translating system specific data coding to meaningful and usable data (the platform where the data will be aggregated does not know that the set of data labeled “Customer No.” is the same as “# Customer” or “ID-Customer”).
3.After cleaning the data, filtering through enormous amounts of data and replicating the application logic to make the data self-describing, the process continues with loading the data in the preferred platform and choosing the visual mode of representation.
In this stage, it can be noted if the background data is very noisy or high quality, as the emerging visual representation will be either hard to read or irrelevant to the strategic objective of the project, or clear and visually engaging.
By implementing key performance indicators along the project and linking them to the project objectives, the increased value will be added in the form of:
- Better quality of the visual representations;
- Fewer project delays;
- Less rework along the way;
- Improved productivity;
- Greater contribution to the visuals’ value;
- Enhanced growth and innovation of the visual representation;
- Easier project assessment.