By the SMU Social Media Team
From the ability to create a successful television series—Netflix’s Stranger Things and iQiyi’s Yanxi Palace—to finding out shoppers’ buying behaviours, and even helping to improve your health, data is now being used for nearly everything.
Today, more than ever, data has become increasingly important in helping businesses and organisations make decisions. According to Mr Rafael J. Barros, senior lecturer at the Singapore Management University School of Information Systems, data analysis has evolved by leaps and bounds in the last five years. “It’s become easier to access. There are now more mature libraries available for multiple platforms, and it’s embedded in our day-to-day business tools like Excel or even Google sheets.”
Making data usable and understandable
Mr Rafael J. Barros, Senior Lecturer at the SMU School of Information Systems
But when it comes to data, it’s not always the case of the more, the merrier. At its core, data is just a bunch of numbers and text, yet it doesn’t exist in a vacuum. It is how we interpret and analyse the data that gives it meaning. So that brings us to the question: Now that you’ve gathered the data, how do you understand it?
Mr Barros sheds some light on the importance of analysing data and how best we should present it to make it useful for others. But before we delve right into how to make data useful, we need to first understand—what is useful data?
According to Mr Barros, data becomes useful through a process known as data cleaning or data wrangling. For a clearer explanation, he refers to the definition from the Consortia Advancing Standards in Research Administration Information (CASRAI) Dictionary: “Data that can be understood and used without additional information. Usable data are delivered in a form that meets the needs of different end-user audiences, is ready for the tasks that the end-user needs to accomplish, and that has been adapted to the end-user’s needs (not the other way around). Usable data have been cleaned, structured, are in a machine-readable format, fully documented, and ready for analysis and interpretation.”
To help ensure that the data is always presented accurately and honestly, Mr Barros recommends following these four tips:
- Is the data correct? What if the report is perfect, but the numbers are not accurate? Then the decision we make based on that information becomes misleading.
- Is the data relevant? This means having a data set that helps us answer the analytical question, which can be a business question that companies need answered.
- Is the data well structured? Humans have a preference for summary tables while computers prefer raw-structured data. We need to transform or clean the human-working data into computer-clean data.
- Is the data complete? Do we have enough data to make a good decision?
And to do so, Mr Barros states that it is necessary to develop the three following skills:
- Critical Thinking. Being able to choose and evaluate what is the best data set available to properly answer an analytical question (either personal or business). This will help with the issues of correctness and relevance of the data.
- Conceptualise and Synthesise. To ensure that you do not get lost in a jungle of data, and to help with the issues of structure and completeness of the data.
- Communication. Using simple visualisations to help others understand the point quickly, and to help share and communicate the analytical findings easily and accurately.
Always tread with caution
However, the easy access to data analysis does not come without worrying dangers, as seen in the manipulation of data in the Facebook-Cambridge Analytica scandal as well as in the proliferation of fake news. “This trend of using data analysis to profile and influence people is worrying, and critical thinking skills are now increasingly crucial in protecting us against it,” cautions Mr Barros.
To prevent the abuse of data analysis, Mr Barros states that it is not about gathering as much data as possible. Instead, one should be focusing on “setting a great analytical question and then collecting the best data available to answer the question”.
This is because at the end of the day, it should not be about “the quantity of the data, but the quality and the purpose of the analysis that matter”.
