View Live/Interactive Dashboard
- Odunlami Zainab
- Samuel Kayode
- Faith Chris-Gbogi
- Happiness Obioma Obi
This project focuses on extracting, cleaning, analyzing, and visualizing Amazon customer reviews for a specific product: MoKo Case for iPad.
The main goal was to evaluate user sentiment, identify recurring themes, and provide actionable recommendations to improve both the product and customer satisfaction.
- Reviews collected using an automated scraper (handled pagination & duplicates).
- Final dataset: 956 unique customer reviews.
- Standardized fields (rating, date, review length).
- Created a unique ID for each review.
- Kept only useful columns.
- Converted review content to string (avoid errors later).
- Converted
datecolumn to datetime → extracted Year and Month.
- Dropped rows with missing dates & handled missing values.
- Ensured clean dataset for analysis.
- Tables: Reviewer_name, Review_id, Review_rating, Review_content, Is_verified, Reviewers, Content_Sentiment, Title_Sentiment, Year.
- Relationships:
- One product → many reviews
- One reviewer → many reviews
- One review → one sentiment
- Review rating distribution
- Sentiment distribution
- Trend of ratings over time (2023, 2024, 2025)
- Verified vs Non-Verified reviews over time
- Reviews by month (all years combined)
- Average rating by month (2025)
- Top reviewers
- Keyword frequency & word content analysis
- Applied VADER/TextBlob sentiment classification to titles & content.
- Built dashboard with:
- KPIs
- Rating distribution
- Sentiment trends
- Word cloud
- Interactive filters
- Scraping Restrictions → CAPTCHAs, IP blocking, throttling (limited review collection).
- Missing/Incomplete Data → some fields reduced accuracy.
- Sentiment Analysis Accuracy → sarcasm/mixed reviews misclassified (e.g., “Not good fit” scored positive).
- Bias in Review Behavior → overrepresentation of very positive/negative customers.
- Limited Scope → focused on one product only (not generalizable).
- Lack of External Context → no sales/returns/competitor data for deeper validation.
- Timeframe Limitations → trends may be affected by external events (marketing, redesigns, seasonality).
- Total Reviews: 956 (since 2022).
- Verified Purchases: 97% (credible, authentic).
- Sentiment Breakdown:
- Positive → 58.16% (~556)
- Neutral → 30.13% (~288)
- Negative → 11.72% (~112)
- Average Rating: 3.3/5 (below Amazon’s 4.0+ competitiveness threshold).
👉 Insight: Large neutral/negative share (42%) drags down competitiveness.
- 5★ Reviews: 78.34% positive → strong satisfaction.
- 4★ Reviews: 68.79% positive, but ~28% neutral → minor gaps.
- 3★ Reviews: 53% positive, ~35% neutral → mixed impressions.
- 1–2★ Reviews: Nearly half neutral/negative → quality/expectation mismatch.
👉 Insight: Customers either love the product or feel indifferent/unsatisfied.
- 2022: 4 reviews (avg 4.8) → strong start.
- 2023: 62 reviews (avg 3.2).
- 2024: 261 reviews (lowest avg 3.0).
- 2025 (YTD): 629 reviews (avg 3.5, slight recovery).
👉 Insight: Sales grew in 2025, but ratings declined post-2022 and only recently started improving.
- Mostly generic accounts (e.g., “Amazon Customer”).
- Few repeat/named reviewers.
👉 Insight: Weak community trust; competitors with engaged/named reviewers may appear more credible.
- Positive (58%) → durability, fit, usability.
- Neutral (30%) → unusually high → unmet expectations.
- Negative (12%) → quality & fit complaints.
👉 Insight: Neutral reviews indicate functionality but lack of “wow factor.”
-
Product Quality & Design Improvements
- Address neutral/negative feedback → fit issues, durability, packaging.
- Run defect theme analysis on 1–3★ reviews.
- Goal: Reduce neutral/negative by 30% in 6 months.
-
Enhance Listing & Marketing
- Add compatibility chart, stress tests, lifestyle images.
- Clarify expectations to reduce confusion.
- Goal: Shift neutrals → positives.
-
Customer Engagement
- Respond to negative reviews publicly.
- Request reviews from satisfied customers (especially recent).
- Encourage photo/video reviews.
- Goal: Raise rating to 4.0+ by end of 2025.
-
Operational & Quality Control
- Audit supplier batches (esp. 2024 dip).
- Monitor return/refund rates.
- Goal: Prevent repeat quality issues.
-
Competitive Strategy
- Benchmark against top competitors → materials, features, pricing.
- Offer differentiation (e.g., durability warranty).
-
Long-Term Brand Building
- Launch “MoKo Care Promise” (1-year replacement warranty).
- Encourage community sharing (Instagram/TikTok).
- Offer bundles (case + screen protector + stylus holder).
- Average Rating → Target 4.0+ by Q4 2025
- Positive Sentiment → 70%+
- Neutral Sentiment → <20%
- Negative Review Ratio → <8%
- Return/Defect Rates → <2%
- Review Volume Growth → maintain strong inflow with better quality
The MoKo iPad Case has strong sales momentum (review surge in 2025) but faces quality & expectation gaps.
Focus on:
- Fixing fit & durability issues
- Clarifying compatibility & product details
- Proactive customer engagement
👉 With these actions, the product can realistically move from 3.3 → 4.0+ average rating, regaining competitiveness in the Amazon marketplace.
- Jupyter Notebook with data extraction and analysis code
- ERD Diagram of the database schema
- Power BI Dashboard with visualizations
- Power BI Dashboard file(PBIX)
- Dataset (CSV)
This project demonstrated a complete data pipeline:
- Data extraction
- Cleaning
- Exploratory Data Analysis
- Sentiment Analysis
- Visualization (Power BI)
While the product is generally well-received, recurring issues (fit, durability) hold it back.
Implementing recommendations will boost satisfaction, reduce negative reviews, and improve sales performance.