Hi ,

​Last year, the Victorian Government released its Digital Strategy 2021-2026, providing a blueprint for the digital transformation of the State of Victoria.

Included in this document is a set of design principles, capturing the Government's vision for what good digital infrastructure looks like. But these principles also provide sound advice to any organisation about to embark on a data science or analytics project.

The principles, along with my own commentary, are as follows:

1. Focus on the customer: Any data science project should focus, front and centre, on the needs of the end user, be they external customers or internal stakeholders.

2. Solve the right problem: Before launching any data science project, understand what the business is actually trying to achieve. Ask "why?"

3. Make it simple: Use data science to simplify existing complex processes, and give preference to simpler solutions.

4. Build on the standard: If a data science solution already exists (for example, in the form of software or model), make use of that first before trying to build something from scratch. And if something does need to be built from scratch, focus on the creation of reusable solutions.

5. Progress over perfection: A "good enough" solution today is worth more than a "perfect" solution one year from today. You can always improve on a "good enough" solution. You can't improve on something that doesn't exist.

6. Trusted by design: Data science solutions should be secure, transparent, ethical and ensure data privacy through all stages of development and delivery.

7. Connect and partner: Wherever possible, in developing data science solutions, make use of the knowledge and experiences of those who have gone before you. This could be through connecting with other data scientists or through the use of resources, such as external data sets or published research.

8. Innovate with purpose: Understand the business value of any data science solutions you develop before commencing development. A $500k solution is wasted on a $100k problem but is invaluable for a $10m problem.

These are worth keeping in mind next time you launch your own data science project.

- Dr Genevieve Hayes

p.s. If you need help in identifying problems suitable for data science solutions, check out my FREE Data Science Project Discovery Guide.