TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

It's Time to Refocus on Data Science and Data Analytics

The coming year is the perfect time to refocus your data science and data analytics efforts to increase revenues and drive core efficiencies.

• By Michael Willock
• December 12, 2019

Multiple times a week, I am contacted by a software startup promising a new platform that will allow me to leverage AI to solve my business challenges and harness new opportunities. Sound familiar? Is this evidence that AI is at peak hype in everyday companies, or is it simply that many companies still struggle to deploy analytics and data science solutions with real bottom-line impact to their business?

For Further Reading: How to Grow a Data Scientist Guided Analytics and the Future of Data Science in the Enterprise How to Cut Data Preparation Time for Visualization Tools

Companies know they should be investing in artificial intelligence and many have spent millions of dollars with no tangible results. Often, under pressure to demonstrate they are investing in "future technology" and not just the same old legacy platforms, enterprises carve out significant funding and spin up data science teams. Unfortunately, these teams often lack connection to the complex dynamics of the business and the projects they tackle do not yield a true ROI. A desire for quick wins can lead data science practitioners to analyze "low-hanging fruit" and, unfortunately, miss out on opportunities for larger business impact.

With signs of economic slowdown and increasing margin pressures across many industries, 2020 will be a year we see companies refocus their data science and data analytics efforts to increase revenues and drive core efficiencies. I can see two big trends emerge from this:

• An increased emphasis on data visualization
• A focus on productionizing data analytics capabilities

Data Visualization

Data visualization has been a red-hot trend for several years, and luckily for us, business intelligence tools continue to improve. Most platforms now have open marketplaces that offer industry-specific utilities to data analysis professionals. As wider audiences of non-data business leaders show interest in data science applications, the use of data visualization to pitch business cases will become more critical. Data visualization and business intelligence tools have been a focus area of my organization's technology learning efforts for this reason. In the financial markets and investments industry, consideration of streaming and unstructured data sources adds significant complexity to this trend. As a data professional, make yourself knowledgeable about patterns and techniques in information design even if you work in an internal IT shop. Effective graphic representation of the data and insights your models produce is critical to convincing a CEO or COO to act. Also, consider borrowing some time from a user experience designer from your company's digital team for that all-important data analytics project you are working on.

Automation and Pipelines

Whether you are using a tried-and-true statistical method for data analysis or are experimenting with the latest machine learning algorithms, your model's potential for impact will be increased by your ability to confidently and repeatedly run it. For this reason, another trend I anticipate is an increased investment in automation and pipelines to support data analytics environments.

As companies look to monetize their data analytics programs, many will place an emphasis on graduating from a research orientation to a production operation. To succeed, companies will adopt data pipeline technologies to support ingestion, scrubbing, aggregation and quality control. If you work with sensitive or regulated data, you have no choice but to advance your data management operation. Regulations relating to personally identifiable information such as the GDPR and CCPA continue to expand.

To address this shift, consider pairing your data scientists with software engineers who have experience with continuous integration and deployment pipelines. If you have yet to consider a cloud strategy for your data assets, now might be the right time. There is an abundance of relevant tools for managing data pipelines that operate natively in cloud environments.

Is 2020 going to be the year you make something more of your data analytics and data science efforts? Increase your probability of success by ensuring you can communicate value with the help of data visualization and run your solutions efficiently via pipelines and automation.