Data science, machine learning, and artificial intelligence are being used to transform businesses and entire industries. These types of projects can provide a company with many benefits, such as improved efficiencies in decision making and in identifying problems and new opportunities. Data science and analytics, in general, are used to uncover patterns, relationships, and provide insight into vast amounts of data. However, it’s not always clear where to begin on analytics projects. Here are some guidelines on how your business can get started.

To Get Started, You Don’t Need a Data Scientist

If your business is just getting started in analytics, rest assured that you do not have to have a skilled and highly educated data scientist on your team to begin visualizing data and getting more insight from it. The reason is that the data visualization software available today makes it easy to ingest the data and begin visualizing it immediately to the point where we can gain new insight into data very quickly. For example, data visualization tools such as Tableau, Microsoft Power BI, MicroStrategy, Domo, and many more make it easy to load data into them and begin exploring right away.

As a start, you might consider placing one of these tools in the hands of someone with some level of analytical skill (probably someone outside of IT, perhaps in customer service, sales, marketing, or operations) who enjoys working with data and has interest in learning a new software tool. Your IT staff may need to help by providing the data or by helping the staff member extract and organize the data. To start, it takes someone who is passionate and can get excited about learning a new software tool and in the process will gain a better understanding of the data.

Explore and Experiment with the Data Using Visualization Tools

Exploring the data and experimenting with it is a great way to start. In fact, you should encourage your newer users of data visualization and analytics to explore and experiment with the data. It’s also helpful, as the staff examines the data, to establish some specific problems and questions that should be answered via the data visualization and analytics. This ensures that users are asking the right questions of the data, such as “What trends are taking place in shipping?” and “Can we identify any seasonality in shipping — perhaps one product as compared to another?” Establishing these specifics up front also ensures that problems and goals for the analysis are very clear from the beginning.

A recent people analytics project asked whether the organization had a diverse workforce. The assumption was that this was true, although the data demonstrated very clearly that diversity was lacking, especially in some departments. Establishing problems and questions up front will help focus and guide the analysis – otherwise, it becomes a slippery slope when your staff begins developing just any old visualization. This way, the data visualizations should be aligned with the overall goals and problems you are trying to solve.

When someone begins an analysis, you can start with an overall problem statement, and then establish questions that address a particular problem. What’s interesting is that when you develop some visualizations that are designed to answer specific questions, there always seem to be additional questions that come up – and that’s to be expected.

Provide Training

Most of the time, training and professional development seem to take a back seat to our regular work. As you begin down the path of implementing analytic solutions, keep in mind that your end users may need some training in this area, even if they already have some analytical skills. At Houghton Mifflin Harcourt, where I lead a team of data analysts, data visualization tools are pretty new to the team, so we have provided both internal and external training so they can gain new skills that can be used immediately to make an impact for our customers.

Take time to train the staff on data visualization tools – either through online courses or microlearning courses. Websites such as Coursera.org, Udemy.com, and Experfy.com offer video courses on all aspects of data analytics and data science – from visualization to learning about specific data-mining algorithms and techniques. Give your staff time to enhance their skills in these areas. Ask them to use those skills in their next visualization or analysis, if possible.

Effective Project Management and Leadership

When a company moves from having only a few people create data visualizations or conduct analyses to a situation where there are bigger challenges and opportunities to explore, effective project management and leadership become paramount. Research suggests that effective leadership and management of an analytics project are going to help set up a project for success. Project managers (PMs) can help keep the team on task but also need to be flexible enough because these types of projects are not always black and white, and may sometimes morph into something that doesn’t look like the original project plan.

PMs should focus on ensuring that everyone is on the same page, and ensure that the goals, problems, and potential solution approaches are all understood by the team. Demonstrating excitement around analytics and becoming a champion for analytics is also important. The staff conducting the analyses need to know that there are leaders and executives that support them and support their work. Remind staff of the importance of the analytics projects and that their work is important and valued.

Be sure to use some type of methodology that will help guide the team during the analytics project. Teams can follow the CRISP-DM methodology, or one of the newer data science methodologies by Microsoft — called the Team Data Science Process (TDSP) — to help guide the project phases and activities. These methodologies are designed to be descriptive rather than overly prescriptive, so they can be easily adapted for your particular business.

Summary

Overall, it should be relatively easy to get started in using data visualization tools to begin building competency in this area. It doesn’t have to be overly complicated to start.

Try these things:

• Give someone who is passionate about data (preferably outside IT) one of the data visualization tools and have them explore the data from their area.

• Give your staff time to explore and experiment with the data. They will inevitably gain deeper understanding of the data and perhaps identify some new insight or problem.

• Develop solid problem statements and questions that will guide the analysis so that you’re not creating visualizations that aren’t aligned with the goals and problems you’re trying to solve.

• Provide training early on in the process. In early phases, it may be easier to have your existing staff skill up rather than hiring a data scientist or data visualization expert right away.

• As you take on more complex analytics projects, effective project management, teamwork, and leadership are critical components. Consider following some methodology for analytics and data science projects, such as CRISP-DM or the Microsoft TDSP.