Data scientists: Here’s how to avoid 3 common mistakes

August 5, 2019

By Alison DeNisco Rayome, TechRepublic

Data scientists remain among the most in-demand tech professionals, but must avoid some common pitfalls to be successful.

Named the best job in America for the last four years, data scientists enjoy high job demand and lucrative salaries, in exchange for their skillsets in big data analysis, machine learning, coding languages, algorithms, and problem assessment. 

Part of the reason for the great demand for data scientists at nearly every company is the rise of the consumer-driven market, said Sri Megha Vujjini, a data scientist at Apexon, a global managed services provider and technology consulting firm. 

“With the power shifting to the consumers, there’s increased competition among companies to be better and there’s pressure to be the option for their customers among the multitude of existing choices,” Vujjini said. “Within this scenario today, there’s not much space for building decisions on experience or a ‘gut feeling’—the market is changing and so are the trends.” 

Companies want to use data and statistics to both improve how they serve their customers, and their own internal problems, Vujjini said. 

“With a rounded skill set and the flexibility to fit in any team (Marketing, Advertising, Pricing, Inventory, E-commerce, etc.), companies naturally want someone who understands their business in a broader sense,” she added. “Data scientists fit perfectly in this scenario—we work as a bridge between information technology and business to come up with solutions built on the internal and external factors and data.”

With so much pressure to help a company perform, certain common mistakes crop up for data scientists. Here are three mistakes data scientists often make, and how to avoid them, according to Vujjini: 

1. Starting with bad data

Data that is incomplete or biased needs to be fixed before it is used, Vujjini said. A company that makes major decisions based off of inaccurate or biased data sets will be setting themselves up for failure. 

Advice: Carefully examine and clean data before using it in any analysis. 

2. Allowing budgetary or timeline restrictions to hurt the work

Data scientists are not always able to use the best tools and methods available due to budgetary and/or timeline restrictions, Vujinni said. 

“The budgetary/timeline restrictions are more common than you think in the industry: Companies want solutions and they naturally want them fast,” she added. “It’s now up to us to balance all the expectations to come up with long-term, quality solutions quickly, and this does result in missing scenarios in the assumptions or spending more time integrating the work being done or even in communicating the progress.” 

Advice: Learn to communicate problems and progress effectively with both technical leads and business leads to keep projects on track. 

3. Not understanding the business 

For younger data scientists in particular, a lack of domain knowledge can be a hindrance to the work being done, Vujjini said. 

“Data scientists must have both the business and technology understanding to be the best as what they do,” she added. 

Advice: Spend time collaborating with colleagues instead of working only on your own to develop a well-rounded understanding of how data is being used across departments. 

“Spending time studying the domain (retail, healthcare, finance, etc.) and interacting with the day-to-day operations of a business helps us understand the data better,” Vujjini said. 

Want to learn more about Apexon? Consult with an expert here.