THE WORK BEHIND
How we crashed our own party and
learned all about our friends’ music tastes.
Not enough time? Scroll down
for the 10-step-version!
#1 Finding that pain point: defining & researching the problem
The first step was to explore and define the problem. After deciding on “group decision making” as a phenomenon, we looked into application areas such as movies, travel destination or restaurants, but eventually decided on our initial idea: music. Then, we dug deeper into previous research and learned a bunch of new stuff about API strategies, global music listening habits and other useful stuff.
#2 Digging into it: analyzing semi-structured interviews
The next step was to find out when, how and why the problem occured. We phoned our friends and family to find out more about when and if this struggle arised, which we then put together into some nice charts and hypotheses to work on.
#3 Putting our (friends’) heads together: focus group session
Based on research and outcome of the interviews, we started to get a sense of what we should gather more info about to come to a solution. A focus group session with seven helpful and smart souls was conducted (in true covid-19 manner – over Zoom) to discuss input, output, connection, music style and much more.
#4 Making it believable: prototyping the front-end
Based on what we found, we started prototyping the front-end of our application using Framer, and manually draw the outlines of a possible music algorithm.
#5 Crashing our own party: test of music algorithm
We asked participants of an upcoming party to send us their playlists, and then manually put together a playlist following our constructed algorithm. The playlist was on all night, and the day after, the participants were asked to share their thoughts on the music played.
#6 Befriending Spotify’s API: Requests and music mixes
Furthermore, we had to get the user data from Spotify to actually produce the playlists. Using Django, we made a first version of the back-end, collecting user information through the Spotify API and putting it together into new playlists.
#7 Have people speaking their mind: think aloud-sessions
Now, it was time to see how good our hi-fi prototype was. A bunch of people got to try out the different features and functionalities, which gave good insights on what we should adjust to make it an actually attractive service.
#8 Wrapping it up: finalizing the product
Eventually, we had collected so much information that we could make the final adjustments and say: hey, this actually turned out pretty good.
#9 Learn your lessons: evaluation of methods and results
Foolishness is making the same mistake twice, right? So, that’s why we tried to see what we could have done better. For instance, the focus group could have been bigger, and the purpose of Groupify was not super obvious to the participants during this activity.
#10 Pop that bottle: end of project
After being done critizing our methods, we felt that it was time to celebrate. The project was over and we all agreed that we deserved some prosecco – and of course, one or two espresso martinis.
Groupify is the result of the project-based course DH2465 Computer Science, Business and Management. An important part of the project has been to include Business, Technology and Design, which you can read more about in the report.
Don’t kill the vibe.