The benefits of the data science function & the role of the CDO within organizations: an interview to Francesca Lazzeri
3 aprile 2019 - Laura Colombo
Francesca Lazzeri, PhD, is a Senior Machine Learning Scientist at Microsoft: she kindly accepted to help us discover the importance of creating a data science function within a company and what are the most crucial steps and attentions to make it healthy and vital.
1) What would you describe as a healthy Data Science function within a company? What would a Chief Data Officer (CDO)’s role be in it?
A healthy Data Science function within a company is a new data-driven framework that organizations need to develop and apply to effectively use data and analytics to change and optimize business processes and perhaps even transform their own industry. The question that arises, then, is – how does that happen? The CDO’s role is crucial to ensure the implementation of a healthy Data Science function. CDOs are responsible for successfully running analytics pilots and build a thriving data science practice. Specifically, they should guarantee that the following key principle are followed:
- Gather the right data. Although companies are collecting petabytes of data, the most important question is: do you have appropriate data for the business problems you are trying to resolve? For example, many organizations that are attempting predictive maintenance have large amount of data available from all sorts of sensors. However, too often, organizations do not have enough data about their failure history and that makes it is very problematic to build predictive maintenance solutions. Models need to be trained on such failure history data to predict future failure events.
- Define a clear vision and purpose around your organization data. Driving innovation requires that organizations have a well-defined and clearly articulated purpose and vision for what they are looking to accomplish. An important step toward this goal is to define a quantifiable business metric and establish how to measure the improvement in the metric with the data science solution (statistically significant A/B test is a common approach to measure this effect).
- Grow a culture of innovation and experimentation. Even where there is a clearly articulated purpose, that alone often doesn’t lead to successful business transformation. CDOs should ensure that the culture of the firm encourages experimentation around the utilization of new data, technologies and ML-based solutions.
2) You talk about embedding Data Science teams to fully engage with a business and adapting the operational backbone of the organization. How radical a change to many non-tech businesses could that be? How necessary is it?
Any data-driven transformation is a radical business transformation, for both technical and non-technical businesses. These business transformations are necessary because they promote the acceleration of business activities, processes, competencies and models to fully leverage the changes and opportunities of digital technologies and their impact in a strategic and prioritized way.
A recommendation that I have specifically for non-tech businesses is to get started by doing a small analytics pilot. This will help exemplify the value of data science solutions in to the rest of the organization. Small pilots are the fastest method to test hypotheses in the real world and promote a data-driven culture across an organization.
Moreover, it is important to have the right AI Stack in place. A modern cloud AI environment will make it easier to collect data, analyze, experiment and finally deploy a data science soluto into production. This sort of capability is becoming a must-have for any organizations, either technical or non-technical.
3) You’ve worked with both AI and machine learning. How do you feel they differ? Is the modern habit among non-scientists of treating them as interchangeable ill-founded or acceptable?
Artificial Intelligence (AI) and Machine Learning (ML) are two very popular buzzwords right now, but they are actually two very distinct concepts. We cannot treat them as interchangeable, as one (ML) is simply a way of achieving the other one (AI).
AI involves machines that can perform tasks that are characteristic of human intelligence. AI can refer to anything from a computer program playing a game of chess, to a voice-recognition system interpreting and responding to speech.
ML is the application of AI based around the notion that we should program our machines, provide them with access to our data and let them learn for themselves. By using a massive amount of data (often called Big Data) the ML algorithm learns how to accomplish goals and improve upon the process.
To learn more about Francesca Lazzeri’s work on Artificial Intelligence and Machine Learning, you can find some of her articles on her Medium account and follow her on Twitter.