Gregorio Tedde

About The Project

My name is Gregorio Tedde and I'm a data scientist based in Sardinia, Italy. After my first degree in Economics, I started my specialization studies on Big Data and since then I've been coding in Python, R and Spark languages in order to deal with machine learning projects, text mining and network analysis every day. I've always been a computer enthusiast since I was a kid, and before growing up with my coding skills, I was already working as a digital graphic designer and as a synthesizer programmer. The latter job is related to another one of my life's passions, which is Music and it actually turned to be a current occupation too. I started taking piano lessons at the age of nine years old, but the real turning point was when my dad gave me a special book as a present for my eleventh birthday. This gift was actually an Italian Deep Purple biography published in 1986, just two years after the successful Mk II reunion. Since then, Jon Lord became my biggest music hero and I began playing organ and keyboards in many local bands across Sardinia, while writing and recording my own songs. I've also started collecting Purple's stuff and everything else related to their family tree. Each record, book or magazine that had the "Deep Purple", "Whitesnake", "Rainbow" tag or other purple-associated names on their cover had become an integral part of my listening and reading routines in the little free time. Off course, my interest in exploring the various relations between all the members that took part in Deep Purple's history and other famous associated acts still grows on me, and this is the reason why I've started wondering: "What if Big Data and Rock Music could deal together?"

I have always admired the amazing job done by Peter Frame (English music journalist) on his Rock Family Trees project. The idea of making the same kind of work in an Integrated Development Environment suddenly began to blow in my mind, so I've started to create my own dataset in April 2019. Several months later, this document turned to be a CSV dataset containing more than 7190 rows, distributed across four columns named after the involved musicians' names, their associated acts, the types of collaboration (bands, albums, singles or live appearances) and the main families (or "universes") to which they belonged (Deep Purple, Whitesnake, Gillan, Black Sabbath and so on). The idea was to add these items inside a Python-generated network and make them act as nodes and neighbors of a huge graph (similar works had already been done on the web with several graphs based on the characters from the Game Of Thrones TV series). Then, each universe would be used to find the main realtionships between each artist included in the original network. So, I started writing the Python classes for building the main graph and, after several tests and many other discovered relations, the work was finally done and the project's scope was reached. Thinking about a possible presentation of the entire analysis, I began to search for solutions on how to deploy the Ipython notebooks I had already wrote, and here we are!

The final network involves (to this day) 5289 nodes and 10825 edges. Each artist's name acts as a specific source, its neighbors (the adjacent nodes, such as band members, producers and other collaborators) are the ones that every Purple expert should actually expect to be and they are all linked together into their associated targets: bands line-ups, friends, collaborations, tour sessionists, etc. Also, many of them act as local gatekeepers between the Deep Purple family members and other solo artists (including their most notorious associated acts). In Network Theory, a node X is defined as a local gatekeeper if there are two of his neighbors (Y and Z) that are not connected by an edge; hence, Y and Z have both links to X, but not to each other. Just to make an example taken from the Deep Purple network, Andy Summers (guitarist from The Police) acts as local gatekeeper between Jon Lord and the node representing The Police, since Summers played guitar in Lord's 1976 Sarabande suite. This kind of relationships emerged for several artists during my analysis, so, in order to avoid graph over-fitting, the numerous solo albums line-ups have been taken into account only if they were particularly determinant in finding stable connections between the involved musicians.

Deep Purple are going to publish their 21st studio album as I'm writing this project's introduction. The main statement of my analysis was to find deep relations between Deep Purple members and other famous artists or producers from the world of Rock Music, that could go further than their respective solo carriers and other well-known side groups. This is the reason why less familiar and "rare" line-ups have also been taken into account, like Deep Purple Mk II playing with Chas Hodges at the Aderbeen Music Hall on March 1971 or the latest Mk VIII sharing the stage together with Jordan Rudess at the Hell & Heaven 2020 Metal Fest. This work was made from a Purple fan to the huge community of other fans around the world, just for affection to one of the greatest rock bands of all time. The entire project has been carried out on a non-profit basis. I really hope you'll enjoy the exploration and if you have any questions or advices, I would really appreciate them. Just contact me here.


GT, April 2020.


Creative Commons License
(CC) 2020 Made by Gregorio Tedde just for analysis purpose.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.