With so much interest in data science, with all the promises of AI, scalable data science, and with the rapidly growing number of trained data scientists, why do we often hear of failed data science projects or projects that remain incomplete with unfulfilled promises?
There are a number of reasons for this, the first of which is a lack of a clear understanding of what data science is: data science is effectively about change – transformational change in an organization to enable a data-driven evidence-based decision-making culture. More technically, I like to define it as follows:
Data Science – a discipline that leverages data, analytics, and technology to affect change by helping make better decisions.
A data-driven paradigm doesn’t necessarily make decision making easier, but it does help make better decisions, and this is the core value proposition of data science.
The second reason is based on a number of common myths of scalable data science, which result in unrealistic expectations, and a lack of understanding of how to embed scalable data science as part of the strategic core of the organization. Let’s address some of these:
- Deep learning is not the panacea for solving all problems, as I’ve previously discussed. You don’t necessarily need the most cutting-edge AI algorithms, running on large, expensive, and powerful infrastructure, to meet your strategic goals. The aim is to first clearly articulate your problem, then determine how data may be used to solve it. You then start with the simplest approach, and only increase complexity when needed. The fundamental aim is to have clear alignment to the strategic goals of the organization, with measurable outcomes to measure success,
- It’s unfortunately not always easy to establish a data science capability with guaranteed and immediate results – there are too many variables at play i.e. people, data, tech, problem scope, and definition – all of which need due consideration, and
- Fear of the unknown, fear of change, and no appetite for risk, will hold your organization back. Educate yourself and your staff in data literacy, so you can comprehend and extract value from data and findings, and increase your analytics maturity. Remember, data science is about exploration, so learn to be comfortable with uncertainty.
Some of these issues stem from inexperienced leadership. This includes insufficient technical leadership that can identify the right problems to solve with the appropriate technical solution, and a lack of an empowered decision-maker who can then support and action the results.
Another common issue is the lack of a data culture that makes it clear to everyone in the organization how data and analytics are used to empower decision making. Without a culture of innovation, collaboration, and trust, and appropriate governance and accountability, it can be challenging to establish and grow your data science capability.
The third common issue relates to technological and data constraints. This includes data scientists being unable to access the data they need (either due to internal silos or complicated disparate data holdings), poor data quality, and inappropriate and unsuitable tools and systems. Sometimes this is due to a blind reliance on specific products, and ad-hoc technology, without a clear understanding of the underlying business problems, and the best technology and tools needed to solve them efficiently, and at scale.
Ultimately, a lack of success in data science is a result of not linking tangible business problems to strategic priorities under the guidance of a suitably qualified and experienced leader, who either directly, or via influence, is able to make decisions.
So, how do you create a successful, sustainable, and scalable data science capability in your organization? Come along to my upcoming talk at ODSC APAC, where I’ll share my 3x T’s for Data Science success in my talk, “How to Establish a Successful, Sustainable and Scalable Data Science and AI Capability within an Organisation“!
About the author/ODSC Speaker, Dr. Alex Antic | Head of Data Science – Australian National University
Alex is a Data Science Leader, Advisor & Educator. He has 17+ years’ experience developed across a number of industries and domains, including Federal and State government, Insurance, Asset Management, Banking (Investment and Retail), Consulting and Academia. Alex has recently been recognised as one of the Top 10 Analytics Leaders in Australia by IAPA (Institute of Analytics Professionals of Australia).