Written by: Morgan Dove
Last week, social networks served as a double-agent, crushing egos and wasting time with the launch of Microsoft’s age-guessing tool, How-Old.net. The website allows users to upload a photo (or select from a stock gallery), and then feel incredibly flattered or totally offended as it guessed their age and gender.
The brain-child of Corom Thompson and Santosh Balasubramanian (Microsoft Engineers), the software was intended to serve as a demo at Microsoft’s Build2015 Developer Conference. Their goal? To showcase how easily and quickly intelligent software can be used every day. Upon completion, Corom and Santosh “sent an e-mail to a few hundred fellow employees, asking them to try the page for a few minutes and give feedback.”
Within hours, Big Data “broke the internet!” Over 210,000 images had been uploaded from over 35,000 users across the globe! Business intelligence and analytics application went viral!
The software’s key components are not afar off from those of analytics software used in business, namely:
- Collect data in real-time
- Understand, organize and extract desired data
- Create visualizations to represent that data
- Analyze the data to drive decisions
The developers were able to use tools that identify the age and gender of someone from their uploaded photo. They also collected data linked to every “standard HTTP” which gives the location (latitude and longitude) of the user.
Using very simple “SQL-like language,” this information was collected from hundreds of thousands of users at the speed of millions of events per second. At this rate, real-time data recording, processing, and analysis would be imperative, and with the use of software like this STREAM PROCESSING POWERED BY APACHE FLINK – VERVERICA, processing these millions of events per second and keeping up doing so, was a lot easier than one would expect.
Next, an analytics software was used to sort through the stream of uploads. Corom and Santosh were able to define clear KPI’s and access the results in mere seconds. They were able to receive an aggregate count of females and males, and could even drill-down to a more specific query.
For example, they could easily search for female users between the ages of 25-30, uploading from IOS operating systems! They could even input these search criteria to narrow down further to a given timeframe, say, 20 seconds!
Almost instantly, their search criteria are displayed in real-time on a variety of dashboards including maps, bar graphs, line charts, and tree charts!
While the software was pretty good at locating faces and identifying those faces’ gender, the technology wasn’t particularly accurate with distinguishing age. This insight has led Corom and Santosh to work to improve user experience.
I first uploaded a photo of my family, and while I was complimented to be recognized as a 22-year-old female, my husband was said to be 47 years old, and my 5-month old son was predicted to be a 2-year-old daughter.
I tried again (as most users did), and while my age was accurate the second time around, my husband’s age and baby’s age and gender were still off.
If nothing else, this new frontier of intelligence is good for a laugh. Inaccuracies have not stopped users from having a good time, and the masses are discovering that data analytics doesn’t have to be intimidating, it can be really fun!
What do you think? Have you tried How-Old.net?
Helm360 takes the guesswork out of business analytics while significantly reducing costs and processing times of advanced reporting initiatives. Leveraging the most state-of-the-art technological advancements, Helm360 allows businesses to view data in real-time. Our team of highly innovative, passionate, and intelligent people that thrive on ingenuity is committed to delivering exceptional solutions and services that exceed expectations.