As a data scientist, one of your biggest struggles in your organization is probably communicating what you do (and what you can do) to the organization outside your data science team. If the departments around your data science team don’t understand how to frame the questions needed for the projects, you aren’t going to get much done. The case for data science literacy is high. As organizations develop data science initiatives, training the rest of the team in data science will be a critical factor in the success of those initiatives. If you aren’t sure where to start, Zachary Brown of Snag has some ideas based on his own experience as a data scientist at S&P working in training and outreach.
[Related Article: Are Successful Data Scientists Hired or Trained?]
Why do we care about literacy?
Continual learning is key to every role within the company. Companies must learn how to leverage new technology. Although you may have a set in stone process for data processing, legacy systems aren’t necessarily the best way to be doing things.
Technical literacy is vital for true collaboration within the organization. Data science teams won’t function best in a vacuum. If a business creates a wall between the mysterious world of data science and the rest of the departments, it can’t take advantage of what data scientists can do. As long as everyone accepts the mysterious status quo, you’ll have a lot of frustration.
If you’re a business stakeholder, you must have an understanding of why the organization should or is migrating to new technologies. Choosing technology stacks is difficult without an understanding of the underlying ecosystem. In no other department would you say, “Well, I don’t understand what’s going on, but someone in that department does.”
You need data science and tech literacy to make informed decisions and improve communication. You want to bring machine learning projects to fruition within a larger organization, so get everyone on the same page.
Training at S&P
S&P has a strategy to bridge the gap between the focus areas that circle data science. Business analysts, data scientists, and software developers, many of whom are fresh graduates, work together to bridge the gap in understanding.
The organization has invested in e-learning modules designed to create communication and understanding, but that’s not the only investment. Literacy investment happens across the entire community, and Brown is currently working on an in-person training session for what data science is and how to interact with it.
Data Literacy Strategies
S&P takes data literacy seriously as an integral part of the entire organization. Brown breaks down what typical literacy strategies look like and how to move from ineffective education to true data literacy.
Traditionally when new tech rolls out, the CEO implements broad training. It’s usually in the form of a module that everyone takes. However, what happens when people go through traditional modules? Unfortunately, the only people who then retain the knowledge from the training is the people actively using it. If your stakeholders aren’t all using the new data science tech, they won’t retain the knowledge from the training.
New Education Strategies
You must transform your education and outreach to be effective. Brown focuses on using open source technologies to transform traditional roles not necessarily associated with the data science team. For example, the common link between data science and business is the business analyst. Creating ways for a business analyst to apply open-source data science tech to the role of business analytics creates a better-informed position.
Training your business analyst to leverage these technologies shifts some of the responsibility for data processing to the business analyst. It breaks down the silos that plague your organization, allowing other stakeholders to bring domain expertise into the process.
Expanding machine learning into software development accomplishes the same type of thing. For Brown and S&P, the ideal environment includes the core data scientist team flanked by business analysts and software developers that now have the same types of skills as the data science team.
To get there, first begin with a few people in your organization that already have some of those skills. They can connect their training with actual use in practice to build a better data science culture. While you can’t expect everyone to 100% adopt the skills of data science, understanding the fundamentals of data science still improves communication around framing problems. The more people you have that can look through the lens of data science, the better.
Implementing Your Data Science Initiatives
There are three components to this rollout:
- Relevance: the data science initiatives must connect to the job you’re targeting. How does looking through the lens of data science transform the tasks and duties of other departments?
- Guidance: traditional, self-paced modules may not be enough. Many people need the expertise of a real person on the team.
- Execution: seeing these workflows in action also helps tie the fundamental ideas of what’s going on to the specific department.
This is a more hands-on approach than traditional training of the past. It can take an organizational direction, starting with the top people and working down, which creates visibility and depth. It could also take a bottom-up approach, allowing for high interest and relevance.
[Related Article: How Developers are Driving Innovation Through Open Source Economy]
Wherever you start, you need to begin your outreach. S&P took a hybrid, grassroots style of approach, mixing open-source courses with instructor-led sessions. Brown or someone from his team would walk through a particular segment and demonstrated how specific models are being used in the company. It increased buy-in and visibility. The connection between the information and the execution within the company was the key to S&P’s success with training.