What is data science and how can data scientists work with your organization? For many organizations, there are a lot of data resources available. Whether it’s your internal member or donor information, program statistics, or open data provided by outside sources. But using that data to its full potential can be a challenge. And many organizations don’t know where to start.
In short, data science is the analysis of data to extract insights and trends. Using data science, an organization can understand problems, explore cause and effect relationships, make predictions about events and behavior, and communicate the results using vizualizations, presentations, and applications.
Today, the ability to collect data and the availability of existing data, make these tools increasingly useful and affordable. Many organizations are focused on using data to meet their goals. Some nonprofits, such as open data orgs, focus solely on collecting and sharing resources. Others are looking to start using it to achieve their missions. The following post is for organizations that are curious about how to begin using data science. From here, you should be able to consult with an expert on how to get the ball rolling.
Whether it’s telling a story or predicting trends, start with these 3 questions:
What problem do you want to solve?
For a mission-driven organization, you may already know what the big problem is. For most organizations, your mission is what drives all your work. So far, you may not have used data to support your goals. Now, you may have a more specific problem that you want to use data to solve.
What will be the outcome of your analysis?
Think about storytelling, making a point, making predictions, or creating software. You may want to communicate some of the key issues that you work to solve. In that case, you might try data visualization. Or, you may want to predict issues in the future. For that, data modeling and machine learning would be an outcome. If you want to empower your team or the public to solve problems, try creating an interactive app.
What data should you use?
Is your data enough, or do you need to look for more? If you collect data already, you might need help wrangling it. If you don’t have data, you can look for open data provided by institutions, governments, and other orgs. If you are collecting data, you will need to consider some ethical questions. Ownership, history, privacy, uses, and bias.
Starting with these questions can be a bit chicken-and-egg. But by focusing on each one at the same time, you can think critically about using data science.
Data science roles
Once you’ve outlined these questions, you can think about who will suit your needs. In this great presentation from Socrata’s 2017 Connect conference, a lot of questions around data science are answered. Merav Yuravlivker, CEO of Data Society, a data science training platform, shared her insights on the field. She reviews a few types of roles to meet your needs. In order from the least to most experience, they are data analyst, data modeler, and data scientist.
A data analyst is the entry-level role for data science. This person would be able to wrangle your data and create basic visualizations. A data modeler can take that information to the next level and use statistical methods to create predictions. A data scientist is the most experienced role. S/he can do all of the above and create custom, comprehensive analysis independently. There are other roles as well, and beyond that, you can build a data science team to meet your needs.
The final outcome will vary based on your needs. But your data scientist(s) should collect, manage, analyze, and visualize your data. Watch the video below for more tips!
Originally published on April 19, 2017.
