Financial Customer Behaviour Dataset

Columns

• age: The age of the loaner, ranging from 19 to 87 years, with a mean of 41.2 years [2].
• job: The type of job the loaner holds, with 'management' and 'blue-collar' being the most common, each representing 21% of the dataset, and 10 other unique job names [2].
• marital: The marital status of the loaner, where 'married' is the most common at 62%, followed by 'single' at 26%, and others [3].
• education: The level of education attained, with 'secondary' being the most frequent at 51%, 'tertiary' at 30%, and others [3].
• default: Indicates whether the loaner was previously in default. 2% of records show a previous default [3].
• balance: The total balance in the loaner's account, ranging from -3313.00 to 71.2k, with a mean of 1.42k [4].
• housing: Indicates whether the loaner has a house, with 57% having housing [4].
• loan: Indicates the loan amount. 15% of records show a loan [5].
• contact: The method of contact for the loaner, primarily 'cellular' at 64% and 'unknown' at 29% [5].
• day: The day of the month when the loan was taken, ranging from 1 to 31, with a mean of 15.9 [5, 6].
• month: The month when the loan was taken, with 'may' being the most common at 31%, followed by 'jul' at 16%, and 10 other months [6].
• duration: The duration of the loan, ranging from 4.00 to 3025.00, with a mean of 264 [6, 7].
• campaign: The number of times a person has taken a loan, ranging from 1 to 50, with a mean of 2.79 [7].
• pdays: This column has values ranging from -1 to 871, with a mean of 39.8. A significant portion (75%) of the data has a value of -1 [8].
• previous: The number of previous contacts, ranging from 0 to 25, with a mean of 0.54. Most records have a value of 0 [8, 9].
• poutcome: The outcome of the previous marketing campaign, with 'unknown' being the most common at 82% [9].
• y: The target variable indicating the loan outcome, with 12% of records showing a positive outcome (true) and 88% showing a negative outcome (false) [9].

Distribution

Usage

• Predictive modelling of loan default or success.
• Customer segmentation based on demographic and financial attributes.
• Analysing the impact of various factors on loan outcomes.
• Evaluating customer behaviour related to borrowing and bank interactions.

Coverage

License

Who Can Use It

• Data Scientists and Machine Learning Engineers: For building and validating credit scoring models.
• Financial Analysts: To understand customer profiles and risk factors in lending.
• Researchers in Economics and Banking: For academic studies on consumer finance and credit behaviour.
• Beginners in Data Analysis: As a readily available dataset for learning data cleaning, splitting, and basic analysis [1].

Dataset Name Suggestions

• Bank Customer Loan Prediction Data
• Credit Scoring for Bank Borrowers
• Financial Customer Behaviour Dataset
• Bank Marketing Customer Data

Attributes