This season, We have analysis to back up my findings and you will we’re going to help you plunge involved with it
A year ago towards Valentine’s, We generated an informal study of your own state out-of Coffee Meets Bagel (otherwise CMB) plus the cliches and fashion We spotted inside the on line profiles female wrote (printed on the another webpages). Although not, I didn’t keeps hard products to give cerdibility to everything i spotted, simply anecdotal musings and you may prominent terms and conditions I seen if you find yourself searching as a consequence of hundreds of users demonstrated.
To begin with, I got to locate ways to have the text message data on the cellular application. The community analysis and you may regional cache was encrypted, thus rather, I took screenshots and you will ran it by way of OCR to find the text. Used to do particular by hand to find out if it might performs, and it also did wonders, but experiencing countless pages by hand duplicating text so you can a keen Yahoo sheet might possibly be monotonous, therefore i must automate which.
The knowledge off CMB is actually tilted in support of the individual’s individual character, therefore the study We mined on users We spotted was tilted into my choice and you can doesn’t portray all profiles
Android enjoys a great automation API entitled MonkeyRunner and an open resource Python variation titled AndroidViewClient, hence welcome full use of the new Python libraries We already had. This try brought in towards a bing sheet, then downloaded so you can good Jupyter laptop computer in which We ran even more Python texts having fun with Pandas, NTLK, and you will Seaborn to help you filter from analysis and you can build brand new graphs less than.
Read More