In a previous post, I gave some advice for those who are interested in a career in data science. One of the suggestions I made was to find a work environment that values and promotes a good data science culture. This is a concept I like to call #DataScienceHappiness, and I believe it’s key for companies that want to get the most out of their data science programs. In order to maximize the return from data science, companies must keep their data scientists happy by taking steps to enable them to be impactful, productive, and feel valued in their jobs. But what exactly is #DataScienceHappiness, and how do you create it?
To promote #DataScienceHappiness, workplaces need to invest in four main areas: talent, process, tools, and culture. I’ll discuss these in more detail below, but an overarching theme to keep in mind is that each of these areas must be tackled at all levels of the company, from the bottom to the top. One of the biggest mistakes companies can make is not having the buy-in of leadership for #DataScienceHappiness, and ultimately, it’s a cultural mindset that requires everyone be onboard in order to work.
What enables a data science team to be successful? Success for the business equals impact. Success for the data scientists equals happiness. Let’s look more closely at these four key areas for creating #DataScienceHappiness.
The first step for setting your data scientists up for success is to be clear about what their role actually is. Data science is a rapidly evolving field, so it’s important to define from the outset what your company’s goals and expectations are. There are many roles a data scientist can play: she can inform product design decisions, implement machine learning, or help you understand your target audience better, among other things. By making it clear what value you’re hoping to get out of your data science program, you’ll make sure you’re attracting the right talent for the role.
Once you have the right candidate for a data scientist position, it’s also key to ensure that you provide them with professional growth opportunities. Again, because data science is evolving every day, practitioners want to continue learning about new developments in the field, so hosting events like lunch-and-learns or hackdays can go a long way towards making your data science team happy.
The second part of promoting #DataScienceHappiness is to invest in the data processes used by your organization. Data science is by nature a very interdisciplinary field, with stakeholders from many different areas of the business, including business executives, product managers, designers, engineers, marketers, and more. A key to success is clearly establishing how data is expected to be used (in which meetings or scenarios and by whom), who is responsible for which aspects of the data, and what standard processes are used for things like success metric definitions, data logging, performance evaluation, and data governance. For example, it is helpful to develop a process which spells out how proper product development should be done, including estimated timelines per phase, to ensure both happiness and success. This avoids scenarios like trying to launch a feature and tacking on expectations to develop good logging or solid data products a few days or weeks before the release. Such processes can be developed for various aspects of data, including: developing good data logging, designing success metrics, and creating a roll-out plan for a new model or feature.
You wouldn’t hand an archeologist a plastic spoon and ask them to go excavate a site; in the same way, data scientists need access to the proper tools in order to do their jobs well. Examples of tools to invest in include data infrastructure, parameters for maintaining sustainable data quality, platforms for self-serve data discovery and querying, metadata and data standardization, and more. If you don’t provide adequate tools for your data science team, they won’t be able to deliver the best results possible. Investing in implementing the proper tools frees up your data science team to be productive and leverages their unique skillset to do higher-level work (model development, insights generation, metric design, performance evaluation, etc.).
This is the most important factor to consider, in my opinion, when creating #DataScienceHappiness. A business must both value data and value data scientists in order to create a positive data culture. Valuing data refers to the idea of having the whole company be invested in things like data quality, data ethics, and data integrity. In order for data scientists to have good material to work with, partners in an organization need to have a healthy respect for data. In the same vein, data scientists need to be valued and viewed as partners, and not as service providers. Making sure data scientists have a seat at the table can go a long way towards helping establish a good data science culture. One way to help promote these mindsets is to have periodic “data bootcamps” or similar sessions that help train employees on how to work with data and data science practitioners. It’s much easier to appreciate work you can understand, so these sessions can be very valuable for the overall business culture.
Data science is an exciting and evolving field, and it can produce real, positive impact for your business—if given an environment to thrive in. Promoting #DataScienceHappiness through the four areas described above can go a long way towards making a data science program the best it can possibly be.
For more details on the concept of #DataScienceHappiness, including specific examples and solutions, you can watch a talk I gave on this subject at the Strata + Hadoop World conference earlier this year.