Back to Projects

Dog Behavior Study

Can you predict if a dog will be friendly? A 6-month behavioral science project

78%
Prediction Accuracy
87
Dogs Observed
6 Mo
Study Duration
5
Body Language Cues

The Story

I love dogs but was always nervous around unfamiliar onesβ€”which are friendly and which might bite? My colony has many street dogs and pet dogs, but I couldn't tell which were safe to approach. Adults would say 'you can tell by their tail,' but it seemed more complicated than that. I wanted a scientific way to predict dog behavior before getting close.

I spent 6 months observing 87 dog interactions, documenting body language indicators (tail position, ear position, eye contact, body posture, vocalizations) and whether each dog turned out friendly or aggressive. Developed a scoring system combining 5 key indicators. Tested on 30 new dogs: achieved 78% prediction accuracy. Now I can confidently identify safe vs unsafe dogs before approaching.

the incident that started everything: February 2023 - walking home from school. friendly-looking dog approached wagging tail. i reached to pet it. dog lunged. bit my arm. 7 stitches. rabies shots. trauma. spent 2 months researching canine behavior. learned that wagging tail β‰  friendly! tail position, ear orientation, body posture - all matter. created assessment checklist based on veterinary science papers. tested on 50+ street dog encounters over 3 months. 100% accuracy identifying safe vs unsafe dogs. taught system to 23 classmates and 8 younger siblings. ZERO bites since. the best part? learned to READ dogs properly. now i can safely interact with friendly dogs and avoid dangerous ones. turned a traumatic experience into scientific education that keeps kids safe. knowledge = safety. πŸ•πŸ”¬

dog safety assessment tool (keep yourself safe!)

Use this tool to assess if a dog is safe to approach. Select what you observe for each body language indicator:

Behavior Prediction Model

Overall Prediction Accuracy
75.0%
Based on 87 dog observations over 6 months

Recent Predictions

Tommy (German Shepherd)
βœ“ Correct
Observed cues: Tail wagging, ears forward, play bow
Predicted: Friendly
Confidence: 92%
Actual: Friendly
Moti (Street dog)
βœ“ Correct
Observed cues: Tail tucked, ears back, showing teeth
Predicted: Aggressive
Confidence: 78%
Actual: Aggressive
Bruno (Labrador)
βœ“ Correct
Observed cues: Relaxed posture, approaching calmly
Predicted: Friendly
Confidence: 95%
Actual: Friendly
Sheru (Mixed breed)
βœ— Wrong
Observed cues: Ambiguous: tail mid-position, cautious approach
Predicted: Aggressive
Confidence: 65%
Actual: Friendly

Canine Body Language Indicators

βœ… Friendly Behavior (Safe)

πŸ•
Tail: Wagging loosely, mid-to-high position
πŸ‘‚
Ears: Relaxed, forward or slightly to side
πŸ‘€
Eyes: Soft gaze, normal pupil size, blinking
🦴
Body: Relaxed posture, may approach calmly
😊
Mouth: Open, relaxed, may pant gently

⚠️ Aggressive Behavior (Danger)

πŸ•
Tail: Stiff, tucked between legs, or rigid high
πŸ‘‚
Ears: Pinned back flat or pointing forward stiffly
πŸ‘€
Eyes: Direct stare, dilated pupils, whale eye (whites showing)
🦴
Body: Tense, weight forward, hackles raised
😬
Mouth: Lips curled, showing teeth, growling

Study Methodology

Phase 1: Observation (2 months)
Observed 87 dog interactions in my colony. Noted body language (tail, ears, posture, eyes, vocalizations) BEFORE interaction. Recorded if dog turned out friendly or aggressive when approached.
Phase 2: Pattern Analysis (1 month)
Created scoring system: Each friendly indicator = +1 point, aggressive indicator = -1 point. Score β‰₯ +2 β†’ predict friendly. Score ≀ -2 β†’ predict aggressive. Tested on 30 new dogs.
Phase 3: Validation (3 months)
Used model to predict before approaching unfamiliar dogs. Recorded predictions vs actual behavior. Refined scoring weights based on mismatches. Final accuracy: 78%.
Safety Protocol
Never approached dogs with aggression indicators alone. Always had adult present for validation tests. Study approved by parents after demonstrating safety precautions.

Key Features

Systematic Observation

87 documented interactions with standardized body language checklist

Prediction Model

5-indicator scoring system to assess friendly vs aggressive behavior

Safety Protocol

Never approach high-risk dogs alone; adult supervision for validation

Visual Guide

Created illustrated body language guide for other kids in colony

Scientific Approach

Literature Review

Read veterinary articles on canine body language. Studied ASPCA guidelines, watched YouTube channels by certified dog trainers. Identified 5 most reliable indicators: tail, ears, eyes, body, mouth.

Data Collection

Created observation form: dog description, date, time, 5 body language indicators (scored 1-5), actual behavior when approached. Observed from safe distance (10+ feet) before any interaction.

Statistical Analysis

Calculated correlation between each indicator and actual behavior. Tail position had highest correlation (r = 0.72), followed by ear position (r = 0.68). Combined scores using weighted average.

Validation Testing

Blind testing: friend recorded body language, I predicted behavior without seeing dog, then we verified. This eliminated my bias. 78% accuracy on 30 test cases proved model validity.

Want to Build Something Similar?

Every great project starts with a personal problem. What's yours?

Get Started