Analysis ideas

Some ideas for analysis

Available Resources:

• Comprehensive data on loops with coverage %, transitions, chord colorings, and modulations based on a million of songs.
• Metadata including YouTube links, genre, country, and timestamp information.
• Possibility to supplement harmonies with MIDI files, and collaborate with platforms like ChordAI or Chordify, as well as tabs and sheet music.
• Audience demographics data (age, country) on who listens to which songs.

Potential Analysis:

• Quick subsequence search capability.
• Analysis of how different chord sequences can continue, similar to that of the Hooktheory TheoryTab DB. It will allow you to drill down the tree of possible progression continuation variants, with probabilities and examples of songs.
• Highlighting of found sequences and repeating loops in the song.
• Analysis and statistics of transitions from one sequence to another.
• Statistics on how often a loop tends to cover the whole song.
• Analyze the relationship between loops, how often they occur together in the same song (compatibility), and perhaps identify clusters of sequences that are more likely to occur together.
• Statistics of possible extra notes in the chords in the sequences (possible 7 / 9 / 11 / 13, bass notes, etc.) - and how it depends on the chord sequence.
• Tonality detection - what key is the song in?
• Chord substitution options within loops and their variations.
• Modulation detection - e.g. find songs with modulations 3 steps up, or how likely are the popular sequences to be in the songs with modulations, or find songs with modulations by a given artist, etc.
• Modulation options - possible transitions and their exact mechanisms, their connection points to the main sequences, etc.
• Bridge creation ideas based on existing bridges in songs with similar chord sequences, popularity of different bridges in different genres, etc.
• Harmonic analysis - detection and naming of chord functions in a given context, and detection of peculiarities of the development of song harmony in a given song.
• Display of maps showing how and where a given chord sequence loop may exit, with probabilities, and allowing to drill down into them.
• Display the statistical popularity of chord progressions in a given song.
• Calculate and display the harmonic uniqueness of songs and specific chords in a given context, allowing you to find the most unique songs or chords.
• Search for strange harmonies, mode deviations, borrowed chords, etc.
• What chord could be played instead in this spot? How can this sequence be altered?
• Enriching the database with data on genres, years, countries, allowing to include corresponding filters in queries and view corresponding output from queries, make corresponding grouping of results, etc.
• Work with unknown parameters. Search and evaluate beautiful or unusual songs by unknown parameters, try to identify these parameters and specific features, analyze the resulting data.
• Clustering moods or themes of songs based on lyrics using OpenAI GPT, trying to find relationships between moods and harmonies.