is the infrastructure of how we collect data and which platform we use to produce results. For example, a company may collect their data via online survey using TypeForm. Then, the data is passed into Amazon Redshift SQL database, then visualized using Tableau.
deals with how the model are created, especially when doing predictive analysis. Mostly this has to be done with coding with R or Python as popular languages. For example, Netflix used machine learning to create recommendation system.
is how to interpret the work of data engineering and data sciences into business sense. For example, the UPS case cited earlier is the example of how to data analysis works. Once learned something from data, we implemented the strategy back into business decision.
1) Data-driven Decision
The companies in this stage have data stored but not cleaned or connected. An easy example here is the street food stall in office area knew that they could sell on average 500 Padthai on weekdays and only 50 on weekends, so they decided to close the shop on weekends. Even though the data was not stored electronically, it still could be used to make good business decision.
2) Integrated Data
Next stage of making decision based on data is to make the data stored properly and integrated. For instance, sales data from marketing department could be combined with profit data from finance department easily. The companies in this stage will dispose of silos data and created the integrated data storage.
3) Real-time Data
Time and tide wait for no man. Having only integrated data would not be useful if the industry is very competitive such as online advertisement. After having the data properly integrated, the aspiration is to have as much updated information as possible. This stage also includes for real-time viewing or analysis of data such as dashboards.
4) Individual Data
Often seen in retail or marketing sector, we do wish to know customers as much as possible and give them the most suitable offer. To be able to do this, we must have individual data of each customers. Even though this might seems easy, the actual implementation is very hard. Try imagine a convenience store which do not have any loyalty program membership, you would not know how often the customer visits or what are the products purchased.
5) Machine Learning
On top of the pyramid, the highest level of data analytics is to use AI (artificial intelligence) or ML (machine learning) to make model adjustment and real-time optimization without human intervention.