Polynumeral was hired by the World Bank to figure out how inexpensive sensor data could be used to improve poverty predictions.
Build a pipeline to make satellite photography crunchable by World Bank analysts. Satellite data is available monthly, and at a fraction of the cost of door-to-door estimates that are only updated once a decade.
The World Bank has proof that Big Data is useful for them. Satellite data improves predictions of poverty by over 10%.
As they figure out their roadmap for the next decade, the Bank has clear evidence that their efforts must include more investment in data science.
A major online service provider hired Polynumeral to analyze referral traffic across the millions of blogs that they service.
Develop an analysis pulling together tens of thousands of gigabytes of data to provide a cross-cutting look at how people move around the web.
Standard Big Data tools would have cost tens of thousands of dollars in computer time to use; instead, Polynumeral’s final analysis ran using only hundreds of dollars and in only a few days.
To understand the complex relationships, Polynumeral designed models to quantify and relate how users move around the Web.
Our client now has clear evidence that their product adds value to their customer base. Their marketing materials feature our result prominently. Their product design decisions have become more coherent as well, since the value of their product has become more evident.
We worked with a consumer health company that produces a wearable device. Their product seemed to work well in testing but took too long to get a good result out in the real world.
Reverse engineer the algorithm and testing process to locate the unexpected slowdown.
After careful analysis, it turned out that the problem was not with the transition out of the lab, but with the underlying measurements of success.
We rewrote the company’s algorithm from scratch. Now results return in seconds.
Complaints dropped, and positive reviews of the connected app rose on both iOS and Android.