By Debashis Tarafdar & Nirmal Palaparthi, Principal Lecturers and Consultants, NUS-ISS
Planning for a Data Analytics Project
One of the most prevalent process re-engineering initiatives of this century is, undoubtedly, digital transformation – that has impacted even the remotest of operations. Every organisation, public or private, is undertaking digital transformation projects to enhance customer-centricity as well as to improve their product and service delivery, at the optimum cost and effort. Together with this, we see a huge move on the part of the consumers to embrace personal digital devices, social media and e-commerce – that perfectly complement the digital transformation initiatives of the organisations that serve these consumers.
As we see the growth of internet-connected consumer electronics and IoT devices that collect large amounts of data in real time - storing, analysing and turning this data into information and knowledge is increasingly becoming a norm rather than an exception. Progressive organisations are using this wealth of data to create actionable insights to serve their customers better, faster and in a proactive manner.
However, challenges still exist for many organisations to manage the process of storing, governing, analysing and acting on the outcomes of such analysis - from technological limitations to process inconsistencies to human skills that drive such initiatives. Core activities need to be taken into consideration, in order to create an effective process for turning data into actionable insights, defined in a 10-Step Data Analytics Project Planning Framework:
The Control Tower View
1. Data analytics projects often suffer from unclear objectives and weak linkages to business goals. A data driven project must be clear, specific and time-bound. This will help develop a strategic project plan and define the right success criteria for the outcome of the project. Hence it is of great importance to prepare by first Defining the Objective.
2. Next Assess the Current Situation. In many organisations, same data resides in multiple silos in different forms. Assess the current available resources including manpower, software, hardware and information. The risks and benefits of the project outcome should be analysed upfront to determine its value and importance.
3. Key concerns about accessibility, security, privacy and policies must be sorted out upfront at the Data Governance step as ownership issues may pose significant challenges during and after the project implementation.
4. The Data Readiness stage is a checkpoint to ensure that the data selected for obtaining, is reliable, credible, relevant and complete. It also allows to check for any biases, irrelevant measures selected and other inconsistencies before one obtains the data.
5. Lack of readiness and maturity of technical resources, capabilities and talent may hinder the timely progress of the project. Consider outcome-based training and dedicated resources to bridge the skill gap. Adopt agile and secure technical platforms to underscore Technical Readiness.
6. Consider how it will be understood and organised before acquiring data, and create a consistent data dictionary. Seek to verify data quality upon collection while Obtaining Data.
7. Be as thorough as you can at the Data Cleaning and Integration stage. Select data that will be used for further analysis and modelling and prepare the data through a thorough data cleaning process. After cleaning the data, it must be integrated into one cohesive format.
8. In Data Visualisation, visualise the cleaned, raw data to identify patterns and trends. These findings can help towards understanding the analytical models to be tested subsequently.
9. Next, choose appropriate modelling techniques and assess the model in the context of your business goals at the Data Modelling stage. Correlation and extrapolation of the data is to be applied for predictive analysis.
10. Last but not least, have a Post Implementation Review, assess the findings and determine recommended actions for the business objectives and reflect on the success of the project. Refer back to the initial purpose of the project and if it has been addressed. Determine next steps with the findings obtained for a full-fledged implementation project.
Enjoy the journey of transforming your organisation to a digital one!
For a more detailed understanding of the Data Analytics Project Planning Framework and associated Project Guide and Tracking Toolkit, email the e-Government Leadership Centre at:
egl-enquiries@nus.edu.sg, or join a Data Analytics course
here.
Fresh from her four-day
DevOps Engineering & Automation class, as part of the
Nucleus 2.0 programme, NCS application consultant, Jeanvia Yeo, shared that "the instructors were very helpful and responsive online. The lecture format were well-balanced with a good mix of assessments, online learning and hands on workshops, with real world applications which I can apply directly to my workplace.” Classmate and fellow NCS application consultant Chin Wei Boon, added that the “online lessons were taught clearly and the instructors where fluent in their teachings.”