Experimental Digital Humanities Project: AI-Assisted Personal Archiving & Analysis

Overview

This project explores how AI can process, analyze, and structure long-term personal documentation—specifically 30 years of journals—using OCR, AI-driven summarization, thematic extraction, and vector-based search. The goal is to develop a methodology for AI-assisted memory reconstruction, cognitive pattern analysis, and thematic evolution tracking.

Key Components

• OCR & Text Processing → Converting handwritten and scanned journals into structured text.
• AI Analysis & Summarization → Extracting themes, concepts, sentiment shifts, and psychological patterns over time.
• Temporal & Semantic Structuring → Organizing insights across different personal eras (Paleolithic, Prologue, Present).
• Vector-Based Exploration → Using Qdrant to map relationships between ideas across decades.
• Praxis Logbook → A living document for reflections on the process, technical hurdles, and AI-assisted self-analysis.

Experimental Angle

The project is positioned within digital humanities, cognitive science, and archival studies, with the potential to:

• Demonstrate how AI can aid in long-term self-reflection and personal knowledge management.
• Develop a framework for AI-assisted autobiographical reconstruction.
• Examine how AI perceives personal history compared to human memory.
• Explore the ethical and psychological implications of AI interpreting personal archives.

Potential Outputs

• A Structured Archive → A digitized, AI-processed collection of journal insights.
• A Research Log (Praxis) → Continuous documentation of findings, challenges, and insights.
• An Interactive Search Tool (Future Idea) → A way to explore personal history through AI-assisted querying.
• A Research Paper or Whitepaper → Summarizing key findings, methodology, and conclusions.

---

PHASE ONE: COLLECTING, COLLATING, CLASSIFYING, AND CATALOGING

Start Date: 17 March 2025

As the past thirty years have passed, a substantial body of physical journals, notes, and various writings has accumulated—now requiring systematic organization, indexing, and transformation into a searchable, structured archive. The methodology for achieving this is multifaceted, involving both the physical documentation process (listing book types, start and end dates, and titles reflective of their contents) and the digital extraction of online journals from various platforms, requiring conversion into usable formats for later analytical workflows.

While it is expected that JSON will serve as a primary data structure in future phases, its utility during this initial phase is contingent on source constraints—the format of exported data being dictated by the properties of the original online platform.

Preliminary Observations & Challenges

A notable shift in documentation habits emerges during this phase. Early volumes tend to have distinct titles inscribed at their outset, a practice that eventually fades. Likewise, inconsistencies in calendar systems introduce an additional challenge in chronological alignment. These temporal fluxations vary by era:

• In the Paleolithic Epoch, fluctuations in record-keeping stem from a wavering commitment to the Ritual of Morning Pages, substance use, and uncertain living arrangements—all contributing to irregularity in documentation.
• In the Prologue Epoch, deviations from the standard Gregorian calendar are deliberate, driven by a fascination with obscure or invented timekeeping systems. The most prominent is a misunderstood and conceptual reworking of The Abysmal Calendar, necessitating post-engineering to translate fictional dates back into Gregorian equivalents for analytical consistency.

A further complication arises from physical degradation across the collection. Many journals suffer from ink bleed, acidic breakdown, and varying paper quality—most of which is not archival grade. This degradation imposes both technical limitations on image-based documentation and a pressing time constraint on data extraction before further loss of legibility occurs.

The Decline of Structure: A Pattern in Handwriting and Cognition

One of the most immediate visual indicators of cognitive and behavioral shifts in the journals is the progressive deterioration of handwriting within each volume. Over the course of nearly every journal, legibility declines toward the later pages, particularly in books situated toward the bottom of the chronological stack.

This pattern suggests a progression from intentional, structured documentation toward a more immediate, impulsive, or fatigued mode of expression. The transition could indicate:

• A decrease in sustained attentiveness, mirroring an inability to maintain focus over long periods.
• A shift from clarity to immediacy, where the act of writing outweighs concerns for legibility or coherence.
• An erosion of the initial intent to maintain structured documentation, implying that the author’s strongest engagement occurs at the outset of new projects, with diminishing investment over time.

This behavioral shift could represent an early marker of an underlying cognitive pattern—one in which the act of beginning holds greater significance than long-term sustainment. Whether this is a symptom of creative momentum, executive function variability, or a psychological aversion to completion remains to be determined, but the pattern itself is clear.