Data science traditionally lives in an ivory tower with quants crunching numbers in a backroom somewhere separated from the business. Because the analytics team works in separate organizations and on different systems, making data science relevant to the business remains the greatest challenge of analytics today.
In the MIT Technology Review article, Data Science and Statistics: Opportunities and Challenges, Devavrat Shah says, “Over the past few decades, we have built infrastructure that can store and process massive amounts of data. However, we still lack the critical ability to seamlessly stitch together various pieces of data to make meaningful predictions that lead to high-impact decisions.”
Data Science and Business Intelligence – All-in-One Systems
Think about how data science makes its way from the data scientist to the business user. The data scientist is the explorer, innovating, writing algorithms, creating new models, working 20 to 25 models at any given time. The data miner experiments, using data mining techniques to identify the top five to seven models. The data analyst validates by checking data quality, ensuring governance, and graduating the top three models to the business. The business analyst looks at valuation, assessing business value, and graduating the top model into production. The business user interacts with the model embedded in a business process, creating value for the organization.
When this entire team works on a single platform, as they do in Yellowfin 7.4, the whole process is accelerated. An average organization can cut days or weeks out of the process that moves the best models into production. The bottom line: all-in-one platforms create more value, faster.
Data Science and the Business Organization
In addition to speed, when the entire analytics team is on the same system, cross-organizational collaboration and understanding happen organically, especially when the system supports collaboration. The system brings the organization together and the team members learn more about the roles and responsibilities of others on the team.
When the business user receives a new or improved model in their business process, they can go back and read through the narrative created in conversations all the way back to the data scientist. When the model goes into production, the data scientist can learn from the narrative provided by the business user. That dialogue is immediately put into action as the data scientist now works to improve the model. This kind of communication only comes when users along the analytics supply chain work on a common system with collaboration built-in.
Yellowfin and Data Science
Is Yellowfin a data science platform? No. Will Yellowfin turn data analysts and business analysts into data scientist? No. But Yellowfin does bring data science models into the business intelligence environment that is already fitted with storytelling and social collaboration.
Consider what a unified platform might mean for your organization. Along with running advanced analytics on data within Yellowfin, analysts can also run predictive models using PMML. The data scientist creates a model and saves it in PMML. The model runs within Yellowfin on actual corporate data. The data analyst can quickly generate visualizations to share the model with the business analyst in a matter of minutes. The business analyst can run Yellowfin machine learning algorithms on the data to identify innovative ways to utilize the data and the best way to visualize the results. Within minutes, the business user has the most understandable view and narrative around the model and can use analytics to make immediate decisions based on the insight.
In this way, Yellowfin BI brings the data scientist down from the ivory tower and give him a seat right next to the business user.
Watch the short video below for a quick view of how you can productionize your data science with PMML.
Want to see how you can use H2O.ai and more?
Book a no-obligation demo where you can see it in action and ask the questions you want.