Five Data Science Predictions for 2016

The data science industry generated many headlines this year, from the U.S. Department of Commerce hiring its first chief data scientist to the National Science Foundation launching four regional data science brain trusts. But now that 2015 is winding down, it’s time to figure out where this buzzworthy industry is headed. Here are some predictions for next year:

Data science will spread to more industries and applications. Eighty-nine percent of corporations believe that not leveraging big data will result in lost market share, a study from Accenture and General Electric found. So as more industries hire data scientists, they’ll be moving away from IT-focused positions and into specialized roles making everyday objects smarter and fine tuning cutting-edge technology.

In fact, the infamous Google Self-Driving Car Project is powered by machine learning that allows autonomous cars to differentiate an exit from a ditch or a child from an adult. Similarly, data science applications will spread to industries like energy forecasting and geopolitics.

The number of data science education programs will increase. According to Indeed.com, the number of data scientist jobs rose 57% in the first quarter of 2015 compared to last year. And with McKinsey & Company predicting a massive shortage of analytics talent by 2018, a boom in data science education opportunities is inevitable.

Data science graduate programs are popping up at institutions like Columbia University, and the number of bootcamp-style programs will increase too. However, bootcamps won’t necessarily turn out qualified candidates, as the skills needed — engineering, statistics, industry knowledge, and creativity — can’t be taught in a few months.

Deep learning techniques will become integral to data science. Deep learning makes it possible to teach systems to recognize images or understand spoken language. It also provides multiple representations of underlying data, generating new ways of predicting and informing behaviors. That’s why this subset of machine learning is a natural addition to data scientists’ toolkits.

Data scientists will use deep learning to automate the process of feature extraction and uncover patterns in data that might have gone unnoticed. Consequently, deep learning tools will become widely available as turnkey solutions. Case in point: In November, Google open sourced its artificial intelligence engine, TensorFlow, which features built-in deep learning support.

Datasets will be bigger, better and more widely available. The amount of data in the world is expected to reach the size of 44 trillion gigabytes by the year 2020, IDC reports. That means a lot more data will be available across a wide array of disciplines.

This will create an “open data” mindset, in which researchers and agencies will share code and data publicly to accelerate learning. At DataScience, we analyze public data from social media to understand the markets we serve. But we also work with expanding open urban data sets to solve bigger problems, like reducing traffic-related fatalities in Los Angeles.

Data science will be adapted to the language of the web. Data scientists rely on programming languages Python and R to create data visualizations, but that could change. That’s because more open-source projects utilize JavaScript, a programming language that is synonymous with the web.

Companies are now open sourcing their JavaScript-reliant components — for instance, Uber has open sourced its mapping component built for React-based applications — and JavaScript’s D3 library makes the creation of interactive data visualizations simpler. More importantly, JavaScript-based data visualizations are easily integrated with web applications, a place where Python and R fall short.

So there you have it: our data science predictions for 2016. Only one thing’s for sure — it’s going to be a big year for data science.

Ian SwansonContributed by: Ian Swanson, co-founder and CEO of DataScience talks about his data science predictions for 2016. Based in Culver City, Calif., DataScience, Inc. combines human intellect with machine-powered analysis to extract information from data that drives real business results. Founded by a team of accomplished entrepreneurs with backgrounds in data science and big data, and experience in such Fortune 500 mainstays as American Express, AOL and Sprint, DataScience traces its lineage to Sometrics, which was acquired by American Express in 2011. DataScience is backed by Whitehart Ventures, Greycroft Partners, Pelion Venture Partners, Crosscut Ventures and TenOneTen.

 

 

 

Download insideAI News: An Insider’s Guide to Apache Spark