Retail · Beginner
Turn a raw export of Jumia listings into an Excel dashboard on pricing, discounts and customer reviews.
You are a data analyst supporting Jumia's marketplace team. Jumia wants to understand how pricing strategies, discounts and customer feedback influence product performance. You have been handed a raw export of home, kitchen and tools listings exactly as it came off the site: prices are text with a currency prefix, some listings carry a price range, review counts are stored as negatives and ratings read "4.5 out of 5". Your job is to clean it, analyse it and build an interactive Excel dashboard that helps Jumia and its sellers make better decisions about pricing, promotions and customer engagement.
Your role
Data Analyst, Jumia marketplace team
Jumia and its sellers set discounts and prices without knowing whether they move customer engagement. Nobody can currently say whether higher discounts bring more reviews, whether well-rated products cost more or less than the rest, which listings are performing well, or which need a different pricing or marketing strategy.
Read the scenario and the dataset overview. Write down the decisions Jumia and its sellers want to make and the questions the dashboard must answer before you open the data.
Keep the raw sheet untouched and build a cleaned copy. Check for missing values, remove duplicate rows, strip the currency prefix and thousands separators from both price columns, turn the discount into a number, make review counts positive whole numbers, and standardise the rating into a numeric value. Record every decision in a cleaning log.
A single table of 115 Jumia product listings with six columns: product name, current price, old price, discount percentage, number of reviews and rating. The export is deliberately raw. Prices are text with a "KSh" prefix and thousands separators, a few listings show a price range rather than a single price, review counts are stored as negative numbers, ratings are text such as "4.5 out of 5", the rating column header is misspelt, and there are missing values and duplicate rows to find.
How to do this project
The scenario
Objectives
Tasks and business questions
Dataset and data dictionaryjumia_products115 listings exactly as exported from Jumia: text prices, negative review counts, ratings as text, duplicates and blanks included. · 115 rows
| Column | Type |
|---|
Add calculated columns: Discount Amount (old price minus current price), a rating category (Poor below 3, Average from 3 to 4, Excellent above 4.5), a discount category (Low below 20%, Medium 20% to 40%, High above 40%) and a price category with thresholds you choose and justify.
Use formulas or a pivot table to establish the baseline figures: average current price, average old price, average discount percentage, average rating, total number of products, total number of reviews, and the most and least expensive products.
Test the relationships in the brief and rank the products. Use scatter or column charts and pivot tables to compare discount against reviews, rating against reviews and price against rating. Produce the top-10 lists and flag the outliers.
Lay out a dashboard sheet with an overview of KPIs (total products, average price, average discount, average rating, total reviews), a product performance section with the three top-10 tables, a trend section with the three relationship charts, and a categories section for rating and discount bands. Add slicers for rating category, discount category and price category, and use conditional formatting to highlight the extremes.
Write the README with your process, key findings, business insights and recommendations. Publish the Dev.to article with screenshots of each stage, push the workbook and README to a GitHub repository, then submit the repository and article links here.
| Description |
|---|
| Example |
|---|
Product | text | Listing title as shown on Jumia. Some titles repeat, which is how duplicate rows are spotted. | Portable Mini Cordless Car Vacuum Cleaner - Blue |
Current price | text (KSh) | Selling price after discount, stored as text with a "KSh" prefix and thousands separator. A few listings show a range. | KSh 2,199 |
old price | text (KSh) | Price before the discount, in the same text format as the current price. | KSh 2,923 |
Discount | text (percentage) | Percentage discount on the listing, stored as text with a percent sign. | 25% |
Review | integer (stored negative) | Number of customer reviews, exported as a negative number. Blank where the listing has no reviews. | -24 |
Ratingd | text (out of 5) | Average customer rating stored as "x out of 5". The header is misspelt in the export. Blank where there are no reviews. | 4.6 out of 5 |
Every price, review and rating converted to a correct number with the cleaning steps documented; duplicates and missing values handled deliberately rather than silently; category thresholds stated and applied consistently; pivot tables and charts that answer the brief's questions; a dashboard a seller can filter with slicers and read unaided; and findings that quote specific figures from the data rather than general statements.
Recommended tools
Skills covered
Excel workbook (.xlsx) containing the original data, the cleaned data, all pivot tables, pivot charts and the interactive dashboard with slicers