Dark Mode
OlistStar Customer Satisfaction Predictor
E-commerce & Online Transactions
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Offers predictive insights into customer satisfaction within the Brazilian e-commerce landscape. Derived from the extensive Olist Brazilian E-Commerce dataset, it features engineered attributes specifically designed to facilitate the prediction of order scores. By aggregating and transforming raw data into meaningful features—such as delivery times, payment methods, and product characteristics—this resource enables researchers and analysts to uncover the drivers of customer experience. It is ideal for building machine learning models aimed at enhancing service quality and understanding consumer sentiment in the retail sector.
Columns
- order_id: The unique identifier for each order.
- score: The customer satisfaction score assigned to the order (ranging from 1 to 5), serving as the target variable for prediction.
- has_more_payment_types: A binary indicator representing whether multiple payment methods were used for the order.
- max_payment_type: The primary or most common payment method chosen by the customer (e.g., credit card, boleto).
- has_sequential: A binary indicator denoting if the order includes sequential numbering.
- has_installments: A binary indicator showing whether the order involves installment payments.
- avg_payment_value: The average monetary value of the payments associated with the order.
- mean_days_purchase_to_approved: The average number of days elapsed between the purchase and the payment approval.
- mean_days_approved_to_carrier: The average number of days from order approval to the shipment being handed over to the carrier.
- mean_days_limit_to_carrier: The average number of days relating to the shipping limit or deadline for the carrier.
- buy_has_work_day: A binary indicator representing whether the purchase occurred on a working day.
- itens: The total count of items included in the order.
- sum_price: The aggregated price of all items within the order.
- sum_freight: The total freight or shipping cost associated with the order.
- sum_same_city: The count of items shipped between a buyer and seller located in the same city.
- sum_same_state: The count of items shipped between a buyer and seller located in the same state.
- mean_distance_km: The average geographical distance in kilometres between the buyer and the seller.
- mean_p_name_lenght: The average length of the product names included in the order.
- mean_p_photos_qty: The average number of photos available for the products in the order.
- mean_p_weight_g: The average weight of the products in grams.
- mean_volume: The average volumetric measurements of the products in the order.
- mean_length_width_ratio: The average ratio of product length to product width.
- mean_cubic_weight: The average cubic weight calculation for the products.
Distribution
The dataset is structured as a tabular CSV file named
order_score_prediction_1.csv with a file size of approximately 8.34 MB. It contains 52,468 unique records, corresponding to individual orders. The data is clean, with 100% valid values across key columns such as order_id and score, ensuring high reliability for analysis.Usage
- Customer Satisfaction Prediction: Develop classification models to predict order scores (1-5) based on logistical and transactional features.
- Logistics Optimisation: Analyse the impact of shipping times (
mean_days_approved_to_carrier) and freight costs on customer sentiment. - E-Commerce Analytics: Investigate the relationship between product attributes (photos, description length) and sales success or satisfaction.
- Payment Behaviour Analysis: Study correlations between payment types (installments, credit cards) and order values.
Coverage
The data covers the e-commerce sector in Brazil. It encompasses a wide range of order characteristics including geographic logistics (distance between buyers and sellers within Brazil), financial details, and temporal data regarding order processing and delivery times.
License
CC BY-SA 4.0
Who Can Use It
- Data Scientists: For training and testing predictive machine learning models.
- E-Commerce Analysts: To derive actionable insights for improving customer service and logistics.
- Academic Researchers: For studying consumer behaviour and satisfaction drivers in online retail.
- Kaggle Enthusiasts: As a clean, feature-rich dataset for exploratory data analysis and competition practice.
Dataset Name Suggestions
- OlistStar Customer Satisfaction Predictor
- Brazilian E-Commerce Order Score Features
- Olist Logistics and Satisfaction Data
- E-Commerce Sentinel: Order Score Dataset
Attributes
Original Data Source: OlistStar Customer Satisfaction Predictor
QUALITYLoading...