Analyzing the Impact of Songs and Backgrounds on Instagram Reel Views

Based on @poetical.whispers's content

---

Intro

In this short analysis, we explore how video production choices, specifically the song used and background visuals influence videoPlayCount performance of Instagram Reels from the creator @poetical.whispers.

The dataset consists of 86 videos, with each video using one of several possible background visuals and one of several possible songs. The variable of interest is videoPlayCount, representing the number of views a Reel receives.

To quantify the effects of these creative choices, we ran a multilinear regression, where each song and background is represented by a dummy variable (1 if used in the video, 0 otherwise).

---

Method

A multilinear regression model was used:

videoPlayCount = β₀ + Σ βᵢ * Songᵢ + Σ βⱼ * Backgroundⱼ + ε

• Dependent variable: videoPlayCount
• Independent variables:
  • Dummy variables for songs
  • Dummy variables for backgrounds
• Reference categories:
  • The song and background not shown in the regression table (implicitly captured by the intercept).

---

Results

Overall Model Fit

• R-squared: 0.042 → Only about 4% of the variation in videoPlayCount is explained by the model.
• Adjusted R-squared: -0.071 → Adjusted for model complexity, the model has very low explanatory power.
• F-statistic p-value: 0.945 → The overall model is not statistically significant.

Song Effects

Top 3 Song Coefficients:

1️⃣ song_name_Slowlydrifting10.wav: +14.06 views (p = 0.589)

2️⃣ song_name_Blinktwice.wav: -20.07 views (p = 0.458)

3️⃣ song_name_Ilookatyou10.wav: -22.36 views (p = 0.388)

Insights:

None of the songs show statistically significant effects (all p-values > 0.05). Slowlydrifting10.wav has the highest positive coefficient, though this is not significant and should be interpreted as exploratory.

Background Effects

Top 3 Background Coefficients:

1️⃣ background_poetic1.mp4: +3.08 views (p = 0.122)

2️⃣ background_poetic2.mp4: -0.11 views (p = 0.997)

3️⃣ background_poetic4.mp4: -10.08 views (p = 0.711)

Insights:

Again, no statistically significant background effects are observed. The coefficients are small and p-values are high, suggesting no reliable pattern in the data regarding backgrounds’ influence on video views.

Summary of Significance

• No statistically significant variables were found in this model.
• The overall model is exploratory and not predictive.

---

Conclusion

Key Takeaways:

✅ song_name_Slowlydrifting10.wav showed the most positive association with videoPlayCount, but the effect is not significant.

🚫 No songs or backgrounds demonstrated statistically significant effects on views.

Cautions:

• The model’s overall explanatory power is very weak.
• Effects should be interpreted as exploratory.
• Future models with more variables (such as posting time, hashtags, engagement history) and larger datasets may improve insights.