← Back to Data

Personal Data Experiments

I pointed AI at 11 years of my behavioral data plus public market signals. Here's what I found.

1.67M+
Rows Analyzed
61
Datasets
11
Years of Data
5+
API Sources

Entertainment, economic, news, and social data sources • Personal exports from streaming, shopping, ads

🎬 What's Trending vs What I'd Watch

Jamie picks only

I love TV and movies—so this was a fun one to analyze. Used AI to cross-reference TMDB trending data, IMDB ratings, Rotten Tomatoes scores, and my viewing patterns to surface what's actually worth watching. ★ = my personal picks.

★Sinners(2025)
IMDB 7.6RT 97%
Max
Mickey 17(2025)
IMDB 6.7RT 77%
Max
★Wicked: For Good(2025)
IMDB 6.7
Theaters
Black Bag(2025)
IMDB 6.7RT 88%
VOD
Frankenstein (del Toro)(2025)
IMDB 7.8RT 90%
Theaters
One Battle After Another(2025)
IMDB 7.5RT 92%
Theaters
Mission: Impossible - Final Reckoning(2025)
IMDB 7.3RT 89%
Theaters
Dune: Part Two(2024)
IMDB 8.4RT 93%
Max
★Wicked(2024)
IMDB 6.9RT 89%
Peacock
Anora(2024)
IMDB 7.4RT 93%
Hulu
The Brutalist(2024)
IMDB 7.3RT 90%
VOD
Conclave(2024)
IMDB 7.4RT 92%
Peacock
The Substance(2024)
IMDB 7.2RT 89%
Mubi
Civil War(2024)
IMDB 7RT 81%
Prime
★Challengers(2024)
IMDB 7RT 88%
Prime
Furiosa(2024)
IMDB 7.5RT 90%
Max
Nosferatu(2024)
IMDB 7.1RT 85%
Theaters
Longlegs(2024)
IMDB 6.6RT 85%
Neon/Hulu
A Complete Unknown(2024)
IMDB 7.3RT 78%
VOD
Heretic(2024)
IMDB 7RT 92%
VOD
Deadpool & Wolverine(2024)
IMDB 7.7RT 79%
Disney+
Inside Out 2(2024)
IMDB 7.6RT 91%
Disney+
★The Fall Guy(2024)
IMDB 7RT 82%
Peacock
Oppenheimer(2023)
IMDB 8.4RT 93%
Peacock
Poor Things(2023)
IMDB 7.9RT 92%
Hulu
Killers of the Flower Moon(2023)
IMDB 7.6RT 93%
Apple TV+
★Barbie(2023)
IMDB 6.8RT 88%
Max

📺 Entertainment Taste Analysis

I analyzed 11 years of viewing history across Netflix and Prime Video, cross-referenced against IMDB/TMDB ratings.

8.18
My Netflix Average
vs
7.0
General Population

Source: Cross-referenced my Netflix viewing history with IMDB ratings database

Netflix = Discovery Mode

New content, prestige TV, binge-worthy series. I'm a curator here—carefully choosing highly-rated content.

Black Mirror 8.7The Queen's Gambit 8.5Ozark 8.5

Prime Video = Comfort Mode

Rewatches, familiar favorites, movies I know by heart. Not discovering—returning to quality comfort.

Pride and Prejudice 7.8Good Will Hunting 8.3Harry Potter series 7.6-8.1

Key insight: The data confirms what I suspected—I have good taste. My average rating is a full point higher than general audiences. I'm not randomly browsing; I use Netflix for new prestige content I'll binge, Prime for comfort films I've seen a dozen times.

📊 Running My Own Ad Experiments

I spent real money running Instagram and LinkedIn ad campaigns to see what the data actually tells you. Spoiler: it's complicated.

Meta/InstagramRich Data

95 campaigns worth of data. Impressions, reach, clicks, cost per result, attribution windows, audience breakdowns.

Verdict: Enough data to actually learn and improve. Doubled down here.

LinkedInSparse Data

23 data points for a single post. Impressions and basic engagement. That's it.

Verdict: Flying blind. The data export gap reflects the optimization gap.

What actually worked:

Timing is everything
Same ad, same audience—completely different results depending on when it runs. Weekends work better for me.
Ads need time to build
Performance tends to improve after a few days. Patience beats panic-optimizing.
Some ads just work forever
Haven't seen much ad decay. Found a few 'always on' winners that keep delivering.
Meta's smart audiences are legit
Advantage+ targeting outperformed my manual attempts. Let the algorithm cook.
Brand alignment wins
Ads that reinforce who I am do better than clever creative that feels off-brand.
You need both acquisition AND retention
Getting new followers means nothing if current ones disengage. Balance both.

The frustrating parts:

  • Results don't always make sense—and that's normal
  • Decoding insights is genuinely hard, even for a data person
  • I feel my clients' pain—I want these platforms to just go ahead and optimize for me! Excited to see how AI helps with this.
  • Comparing to past campaigns? Industry benchmarks? Good luck doing that manually.

Note: These are my personal observations from running my own small campaigns. Your mileage may vary. Not professional advertising advice.

🔬 Other Analysis

What else I've looked at with this data:

🛒

Purchasing Data

Analyzed Amazon and Target order history. Verdict: I'm a pretty predictable shopper. Not much exciting here.

Status: Analyzed, meh results
📈

Cross-Platform Ad Patterns

Stumbled onto some interesting patterns comparing my behavior across platforms. Might be something worth exploring further.

Status: Exploring—more to come