MIT (Massachusetts Institute of Technology) has created a tool which can predict how many likes your Instagram pic will receive. The ‘Popularity Demo’ determines what makes an image popular through two factors: what’s in the picture and the social context. The clever tool was invented by Aditya Kholsa (MIT), Atish Das Sarma (eBay Research Labs) and Raffay Hamid (DigitalGlobe) by “using a dataset of about 2.3 million images from Flickr” the research team “can reliably predict the noramlized view count of images … using both image content and social cues.”
Aspects that the team think contribute to how many times someone double taps are “color, gradients, deep learning features and the set of objects present, as well as the importance of various social cues such as number of friends or number of photos uploaded that lead to high or low popularity of images”. Their research paper also states that the most appealing colours on Instagram are medium pinks, lime greens, dark purples, bright reds, yellow and sky blue whilst dark greens, medium blues and greys are the least popular. Their research shows that the most popular Instagram items to get likes are bras, revolvers, mini skirts, maillots, bikinis and weirdly – teacups. This seems plausible – scantily clad women were popular images even before the invention of Instagram! On the other side of the spectrum the most negative images were plungers, laptops and spatulas – so no more postings pics of these then! The middle ground of popular items included cheetahs, giant pandas, basketballs, llamas, plows and ladybugs.
The ‘Popularity Demo’ asks you to upload a pic (either from your hard drive or a URL) and then gives you a rating based on all the factors above.
What do you think Collegiettes? Is this the best thing ever or a bit of a waste of time? How many predicted likes did you get? Comment below!
Sources
http://people.csail.mit.edu/khosla/papers/www2014_khosla.pdf