When used correctly, LLMs have been a blessing for UFO Data Science, and when given the right data and prompts. See prior work posted to this sub by VerbaGPT, deckerRTM, and Pristine_Ad3669.

Last week, I switched from ChatGPT to Claude Code and the difference has been night and day (and not just because Claude generated the graphs in dark mode!)

I used up-to-date data from just MUFON and NUFORC, which I can't share here because it's proprietary. But don't worry, I did open-source a project with 331K reports up to 2023. It uses the old UPDB.app PostgreSQL database, available on the Wayback Machine, with 12 sources including MUFON and NUFORC.

Here's the initial prompt I used for my analysis:

Use Natural Language Processing and sentiment analysis to generate insights about what emotions people feel when they see a UFO, by analyzing the CSV datasets in this directory.

Edit: The model I used was Claude Code v2.1.50 Sonnet 4.6 · Claude Pro

Below is the generated output.

UFO Witness Emotion Analysis — NLP & Sentiment Report

Dataset: MUFON + NUFORC | Reports analysed: 288,037 | Generated: 2026-02-22

Dashboard

UFO Emotion Dashboard

Overall Sentiment Distribution

Mean VADER compound score: 0.2693  (range: −1 = most negative, +1 = most positive)

Emotion Category Prevalence

% of reports containing at least one emotion-matched word

Emotion Word Share

Share of all matched emotion tokens

Mean VADER Compound Score by Source

Emotion ↔ Sentiment Correlation

Pearson r of each emotion count vs. VADER compound score

Dominant Emotion per Report

Word Cloud

UFO Witness Vocabulary Word Cloud

Sentiment Score Distributions by Emotion

VADER Sentiment KDE by Emotion

Emotion Radar — Overall vs MUFON vs NUFORC

Emotion Prevalence Radar Chart

Key Insights

1. Curiosity & Awe Dominate — "Curiosity" appears in 40.5% of all reports, closely followed by "Calm / Peaceful" at 24.7%. Witnessing a UFO primarily triggers a drive to observe and make sense of what is seen.
2. Mixed Sentiment — Wonder Edges Out Fear — Positive: 62.7% | Neutral: 9.0% | Negative: 28.4%. Despite the strangeness of sightings, more witnesses describe the experience in positive or neutral terms than in fearful or negative ones.
3. Dominant Single Emotion — When one emotion dominates a report, it is most commonly "Curiosity", followed by "Confusion / Disbelief". This suggests the first reaction is usually intellectual (trying to identify the object) before emotional.
4. Fear Correlates Most Negatively with Sentiment — Reports heavy in fear-language score lowest on VADER, confirming that fear-laden accounts use strongly negative vocabulary (terrified, horror, panic).
5. Awe and Calm Correlate Positively with Sentiment — Reports describing beauty, silence, or transcendence use more positive language, suggesting a subset of witnesses find the experience spiritually uplifting.
6. MUFON vs NUFORC Tone — MUFON reports (mean 0.2744) tend to be slightly more positive than NUFORC reports (0.2649), likely reflecting differences in reporting style (investigator-mediated vs. self-submitted).