You may have been doing basic data analytics without even knowing. For example, when you compared the benefits of owning an IPhone versus a Samsung Galaxy (or any other preferable brand) based on information from different sources, whether you realized it or not, you have ultimately gathered, inspected, cleansed, transformed and analyzed data that helped guide your decision making process of acquiring a new phone – and this is exactly what data analytics is about.
Businesses are also known for their extensive use of data analytics to guide decision making. E-commerce giants such as Amazon and Alibaba often study data to help predict when and where clients might need certain products. This serves to help cut costs associated to shipping, warehousing, labour, among others.
Governments on the other hand, may use data analytics to support policy, budgetary, and operational decisions. For example, government agencies will often use data analytics to forecast natural occurrences that may impact citizen’s lives. Chiefly, agencies such as the National Oceanic and Atmospheric Administration (NOAA) in the US, assist the government in predicting a hurricane’s trajectory by looking at current and historical data. Such information, is used to allocate appropriate resources to affected areas, ultimately resulting in lives saved and decreased damage costs.
Data Analytics is the study of data (often through statistics) to better understand a topic and/ or support individuals, businesses and governments throughout their decision-making process. Thanks to this science, information can be analyzed in several different ways in order to expose patterns and trends, resulting on more efficient business practices and better decision making.
Data Analytics has been traditionally divided into 4 segments:
In order to make data driven decisions, data analytics uses the following process to guide the decision-making process:
Raw Data – Refers to information that has not been changed, processed or filtered to the point where it can be used for analysis, presentation and decision making. For example, every action taken throughout the process of purchasing something online creates data. Without connecting the dots (raw data) of some or all the process, that data on its own has no meaning.
Unstructured data – Refers to data that does not conform to a structure that allows for easier data processing and analysis. Pictures, e-mails, sounds, video and text without the right structure will often fall under this segment. For example, a company has pictures of multiple consumers purchasing a variety of products, the picture may provide information regarding demographics, age, purchased product, amount spent among other areas. However, pictures are not structured in a pre-defined manner that facilitates data analysis, making it difficult to dig into the information without spending significant resource.
Structured data – Refers to data that has been compiled in a way that facilitates the data modelling process, which is often found in text format and easily examined. An example is provided below:
Note: With the advances in technology driven by the Fourth Industrial revolution, data analytics derived from pictures, videos & audios are starting to become more commercially viable due to decreased costs of data storage and advancements in tools such as Amazon Rekognition & Google Vision AI.
Data Cleansing – The process of finding, fixing or removing inaccurate or corrupted data to support the data analytics process. An example can be seeing below:
Data Report – A data driven analysis that has defined, created and optimized information in an accurate manner to help support the decision-making process.
Ad Hoc Data Analysis – Data driven analysis with the aim of digging into a segment of a dataset or prior results to answer specific questions that may have been raised. For example, the data report above can be used by the government to determine key factors that may lead certain regions to have a higher rate of homicide.
Data Filtering – Closely related to ad hoc data analysis, data filtering looks to set filters (conditions) to dig further into already available data, focusing on areas relevant to the author or decision maker.
Data Alert – A set of pre-assigned specifications/limits (whether fixed, formula driven or otherwise) implemented in reports or data tool(s) that raises alerts when the data goes outside defined ranges or is expected to do so based on forecasted information.
Pattern Recognition – The discovery of patterns and irregularities in the data driven by man-made rules, heuristics, machine learning or “Artificial Intelligence” to help drive decision making (E.g. data clustering, anomaly analysis)
Data Analytics Process Automation aims to automate repetitive tasks associated with data analysis to help decrease costs, improve the bottom-line, ensure information arrives to the right people at the right time and frees employees’ time to do more value-added tasks. Data analytics process automation can help structure data, cleanse it, create the reports automatically, recognize patterns, and set alerts among many other areas.
Training & Workshop – Ensures that teams are well equipped to do quality data analytics work. It can also serve to help operational teams manage data analytics tools and automated systems, leading to long-term organizational success. For an example of a data analytics and visualization report, please refer to ‘Canadian School Expenditure Report’