This is a continuation of my Data as Experience Series. The last post was on the Power User role.

The Data Scientist

At first glance, you might ask why even consider a data scientist.  From the typical analytics perspective of, “What reports should I write?”  this role demands nothing.  But when you think about what they do, you see that you still need to consider their experience and plan for it accordingly.  But first, let’s look at the typical day in the life of a scientist.   Many thanks to Matthew Mayo at KDnuggets for insight from his various contributors.

Data scientists have a variety of needs, none of which come with pre-defined reports given that they are meant to live in the data and suss out understandings from it as they create models and even train AI’s.

Their Tools:

• Data extraction tools
• Excel
• R
• SAS
• Python
• SQL to a DB
• Various Big Data products like Hive, Hadoop HUE, etc.
• Various BI analytics tools (think Tableau, MicroStrategy, PowerBI, etc.)

What they do:

The Future of Big Data

With some guidance, you can craft a data platform that is right for your organization’s needs and gets the most return from your data capital.

Get the Guide

When you look at the day to day, you see some common themes occurring that can be broken down accordingly:

1. Communication and coordination with their teams
2. Understanding of business and business needs
3. Getting at the data and ensuring the data can be trusted
4. Creating models to gain insight or train an AI
   1. Includes time to test and iterate on the model
5. Create visualizations to show insight gained to the business
6. Research regarding how this still nascent field continues to mature

What An Analytics Team Should Do

As you can see, a typical data scientist has far more technical chops than even a power user.  They can do a lot themselves and given their unique needs, will demand from you.   That means you need to focus on three things that will help them.  Frankly, if you do these two things right, they will love you forever:

1. Create well governed and curated data.  You can do so by pulling in data that’s trusted to begin with. You can also do it by using BI governance tools to manage the data and make it available as trusted, audited, and understood.
2. Provide the tools necessary to pull the data out of the source repositories easily and quickly.  They need it to run through their own tools, create models, etc.  You can’t help them do this job so just let them pull what they need and get out of the way
3. Give them a good visualization tool.  This has certain implications
   1. You have a good visualization tool
   2. You have given the data scientist enough training to use the tool
   3. The data scientist has the right authority to use the tool to its fullest extent. In other word, make them a report writer with the authority to access a range of data sources including their own recently created data.

Bottom Line

Don’t treat the data scientist like a typical user.  Give them what they need and then get out of the way and let them do what they need to do.  They will pull, analyze, and report on all this themselves.

Up Next

Executives remain a top priority of users and continue to need daily information and unique insights.