Could APIs provide advanced analytics for the masses?

These building blocks of the app economy might open new markets for analytics

By Alison Bolen, SAS Insights Editor

You don’t have to be a developer to understand the usefulness of APIs (short for application programming interfaces). In fact, most of us interact with APIs every day and don’t even realize it.

If you’ve printed photos from an app or refilled prescriptions from your phone you’ve used an API. If you’ve shared an article to Twitter from the article page, you’ve used an API.

At work, if you view customer data on a Google map, you’ve likely used an API. If you can compare social media statistics from multiple channels inside a single application, that’s happening through APIs too.

APIs have been called the building blocks of the app economy. They are portable pieces of code that can be easily combined and stacked together to enhance other applications or websites.

An API for data scientists

So what if you’re a data scientist, and you want to add forecasting or scoring capabilities to your purchasing system or inventory program? In the past, only the simplest of tasks were available as APIs, but today more companies are figuring out how to turn increasingly complex tasks – and combinations of tasks – into APIs.

“With SAS^® Viya^® we can provide REST APIs for data scientists to embed analytics into their ongoing projects,” says Wayne Thompson, Chief Data Scientist at SAS. “You can easily go out to the cloud, grab an API for prediction, segmentation, forecasting, scoring or sentiment analysis, and plug it into an application that you’re building. The APIs help free up valuable time so you can focus on application building instead of writing, testing, assessing and algorithms.”

This self-service functionality will make it easy to add analytical capability to basic BI reports or enterprise applications, providing a huge advantage for programmers and data scientists. “If I want to understand supply and demand and I want to build a site that monitors that, I can focus on data ingestion, JavaScript and HTML,” explains Thompson. “For the analytics piece, all I have to do is drop that forecasting API in there and it gives the answers.”

 For companies that want to use their own proprietary models, APIs can make it easier to share those models in house with other departments and divisions. Within a large organization, one developer might create an attrition model that would then be easily embedded into CRM systems globally.

 REST APIs, which provide a standard method for communicating data from one application to another, are also making it easy to call SAS from other applications where data scientists are working. With SAS Viya, for example, data scientists can easily connect to SAS from Python, Lua or Java, to name a few.

Inspiring innovation with analytics

By providing APIs, many companies have found that opening their code to others allows for more innovation, and suddenly the original application or service gets used in ways the originator would not have imaged.

“When traditionally nonsoftware companies like Walgreens and Nike opened their data and services via APIs, they attracted entirely new business partners and innovative app developers,” says David Biesack, Principal API Architect at SAS. “APIs enable innovation and give people the opportunity to build mashups of multiple services and create entirely new things.”

The result is that other companies flocked to build functionality for mobile phone users that Walgreens, Nike and other companies themselves could not provide.  APIs become a huge force multiplier.

Likewise, many automakers are now opening up their APIs, resulting in voice-activated functionality and other dashboard applications that the manufacturers themselves never envisioned.

Will the analytics world see similar results, as entrepreneurial thinkers come up with new ways to add analytics into systems that no one ever thought of before?

Just as APIs have made mapping technology ubiquitous and voice activation more common, they could also make advanced analytics more prevalent, opening up entire new markets to use the predictive capabilities of analytics – and providing answers to segments of users who’ve never used analytics before.