Can You Hear the Difference Between a Real Scientific Trend and a Data Artifact? 🎧
Can you hear the difference between a real scientific trend and a data artifact? 🎧
I recently collaborated with Lutz Bornmann on a project for the LSE Impact Blog (London School of Economics & Political Science) to investigate whether scientific progress is actually losing its disruptive edge.
For this piece, I wrote the article, produced an animated video, and created a custom data sonification mapping 34 years of citation data across 30+ million publications.

Watch the Animated Sonification Video
The Data Sonification Process
To turn decades of scientometric indices into an intuitive acoustic experience, I developed a custom data processing workflow:
-
Data Mapping (Python & Scale Quantization):
I wrote a Python script to map annual average Consolidation-Disruption Index (CDI) values directly onto pitches drawn from a 4-octave C Dorian harmonic scale. -
Dynamic Velocity & Density:
Publication volume was mapped to note velocity (loudness), while the arpeggio density was programmed to dynamically respond to rising or falling CDI trends in real time. -
Audio Production (Ableton Live 12):
I exported the generated MIDI data and produced the final spatial tracks in Ableton Live 12 using custom synthesizer patches, warm analog filtering, and ambient room reverb.
The Result: What Does Scientific Progress Sound Like?
The sonification reveals a striking acoustic contrast:
- Without Thresholds (Raw Artifact): The pitch steadily slides down over time, representing an apparent decline in scientific disruption.
- With Citation Filters (Filtered Trend): The pitch steadily climbs, proving that when proper citation thresholds are applied, scientific disruption continues to grow.
Read the Full Article
Explore the complete methodology and scientometric analysis on the LSE Impact Blog:
🔗 LSE Impact Blog: Read the full article on LSE
🎥 YouTube: Watch the video on YouTube
#DataSonification #ScienceCommunication #AbletonLive #Python #DataScience #CreativeCoding #Scientometrics