Correlations¶

Correlations are the heart of Intestigator. They show you which foods or drinks tend to precede your symptoms.

What You'll See¶

Each correlation shows something like:

Bloating often follows Dairy 5 of 8 dairy exposures Typical delay: ~2 hours Based on 5 examples

This means: Out of the 8 times you logged dairy, bloating followed 5 of those times within a few hours. The typical gap between eating dairy and feeling bloated was about 2 hours.

How Correlations Are Calculated¶

Intestigator looks at your logs and asks: "When this person logged [symptom], what did they eat in the hours before?"

It then counts how often each food category appears before each symptom type. If a pattern shows up enough times, it becomes a correlation.

Time Windows¶

You can adjust the time window Intestigator uses:

Try different windows if you suspect your symptoms have unusual timing.

Analysis Modes¶

You can analyze correlations using three different modes:

• Ingredients — Analyzes specific items you've added to your meals (this is the default)
• Tags — Groups foods by category (dairy, gluten, caffeine, etc.)
• Descriptions — Looks for patterns in your meal description text

Choosing the Right Mode¶

Ingredients works best when you've been consistently adding individual items to your food entries. It gives you the most precise results because it tracks exactly what you ate.

Tags is helpful when you want a broader view. If you suspect "something with dairy" causes issues but aren't sure which specific dairy product, tag analysis groups everything together.

Descriptions analyzes the names you give your entries (like "Pepperoni Pizza" or "Morning Smoothie"). It groups similar descriptions together and looks for patterns, which can surface connections when you haven't added tags or ingredients.

Automatic Fallback¶

When you first open correlations, Intestigator tries to find meaningful patterns automatically:

1. It starts with Ingredients (the most specific)
2. If no patterns emerge, it falls back to Tags (broader categories)
3. If still nothing, it tries Descriptions (entry names)

So if you notice the mode has changed from what you expected, it's because the app found better results in a different mode. You can always switch manually using the mode picker.

Viewing Examples¶

Tap any correlation to see the actual entries that created it:

Example 1 Ate: Cheese pizza (12:30 PM) Felt: Bloating (3:15 PM) Delay: 2 hours 45 minutes

This lets you verify the pattern makes sense and recall specific incidents.

Minimum Requirements¶

Correlations only appear when there's enough data: - At least 3 times the symptom followed the food - At least 5 total exposures to that food category

This prevents false patterns from small sample sizes.

What Correlations Don't Mean¶

A correlation is not proof of cause and effect. It means "these two things often happen together in your data." Possible explanations: - The food actually triggers the symptom - Something else you eat with that food triggers it - The timing is coincidental - A third factor affects both

Use correlations as clues to investigate, not as diagnoses.

Next: Learn about Trends.