Sentiment Analysis of AliExpress Product Reviews in Spreadsheets and Its Impact on Product Improvement-OK Spreadsheet-The best and new spreadsheet

THE NORTH FACE Down jacket（28 styles）-0234

$30.42

AirPods 3-0102

$7.64

Jordan 4 Retro (40 color)-0001

$34.72

New Balance 9060 series (20 style)-0011

$38.89

2024 New Jordan 4 (12 colors)-0010

$38.89

Rolex watches (3 qualities 14 styles)-0396

$52.78

GOYARD Fashion bags-0182

$9.58

Balenciaga Triples (7 styles) -0515

$83.33

UGG CLASSIC MINI（27 Style）-0013

$25

Louis Vuitton fashion bag-0185

$48.61

Louis Vuitton LV belt（36 style）-0246

$11.81

Apple iPhone 17 Pro Max 512GB-4745

$180.42

Apple AirPods 4 (with active noise cancellation) with wireless charging case (USB-C)-4744

$16.53

Canada Goose new high-end down jacket short shiny-4313

$133.33

AMIRI MA-1-0024

$72.22

Burberry Down jacket-0827

$38.89

UGG Tazz comfortable soft fashion wear height increase warm-4500

$34.72

Jordan 4 anti-slip wear-resistant mid-top retro basketball shoes-4241

$123.33

Apple Air Pods Max (USB-C)-4841

$97.08

Apple AirPods Pro 3 Bluetooth Earphones with MagSafe Charging Case (USB-C)-4743

$17.92

Introduction

In the era of e-commerce, customer feedback plays a crucial role in shaping product quality and user experience. This article explores how AliExpress product reviews can be analyzed using sentiment classification techniques in Google SpreadsheetsMicrosoft Excel, combined with text mining toolssentiment analysis algorithms. The goal is to extract actionable insights to guide product improvements.

Methodology

3-Step Analytical Process:

Data Import: Export product reviews (text ratings, comments) into structured spreadsheets
Sentiment Classification:
- Use =GOOGLETRANSLATE() to convert multilingual reviews
- Apply lexical-based analysis (AFINN, VADER) via scripts
- Categorize as Positive/Neutral/Negative
Keyword Extraction:
- TF-IDF analysis using built-in functions
- N-gram frequency calculation (for phrases like "slow delivery")

Product Improvement Applications

Quality Enhancements

Negative reviews mentioning "battery life" (frequency score: 0.78) indicate need for power optimization in electronics.

Feature Prioritization

38% positive reviews highlight "easy installation" feature, confirming design effectiveness.

Packaging Improvements

"damaged box" appears in 12% negative reviews, suggesting materials upgrade.

Implementation Notes

⚠ Challenges:

Colloquial expressions (e.g., Chinese "不错") may require manual sentiment calibration
Statistical formulas have 75-85% accuracy compared to ML models

Conclusion

This low-code analysis approach

Quickly identify top complaint categories (logistics 35%, material 27%)

Validate product modification decisions with data

Track sentiment trends after improvements (see Case Study A)

For automated implementation, consider combining with Apps Script or Power Query for batch processing of 1000+ reviews.

```