Adult Income Prediction Dataset

Columns

• age: The age of the individual, ranging from 17 to 90 years, with a mean of 38.7 years.
• workclass: Categorises individuals based on the nature of their employment. Common categories include 'Private' (69%) and 'Self-emp-not-inc' (8%). There are 9 unique workclass types.
• final_weight: A numerical weighting factor associated with each individual, with values ranging from approximately 12,300 to 1.49 million, a mean of 190,000, and a standard deviation of 106,000.
• education: Describes the type of education received by a person. 'HS-grad' is the most common (32%), followed by 'Some-college' (22%). There are 16 unique education types.
• education-num: A numerical representation indicating the education level, with values from 1 to 16, a mean of 10.1, and a standard deviation of 2.57.
• marital_status: Indicates the marriage status of an individual. 'Married-civ-spouse' is the most frequent (46%), followed by 'Never-married' (33%). The dataset contains 7 unique marital statuses.
• occupation: Specifies the type of occupation held by an individual. 'Prof-specialty' and 'Exec-managerial' are the most common, both at 13%. There are 15 unique occupation types.
• relationship: Refers to the individual's role within a household or family structure. 'Husband' is the most common role (40%), followed by 'Not-in-family' (26%). There are 6 unique relationship types.
• race: Categorises individuals based on their racial or ethnic backgrounds. 'White' represents 86% of the dataset, with 'Black' at 10%. There are 5 unique racial categories.
• sex: Refers to the biological sex of an individual. 'Male' accounts for 67% and 'Female' for 33%. There are 2 unique categories.
• capital_gain: A numerical attribute representing capital gains, with a mean of approximately 1,090 and a standard deviation of 7,520. Most entries (34,789 out of 36,600) have a capital gain of 0.
• capital_loss: A numerical attribute representing capital losses, with a mean of 85.9 and a standard deviation of 400. Most entries (34,958 out of 36,600) have a capital loss of 0.
• hours_per_week: The number of hours worked per week, ranging from 1 to 99, with a mean of 40.4 and a standard deviation of 12.4. The most frequent entry is 40 hours per week (18,082 individuals).
• native_country: The country of origin for the individual. 'United-States' is the most common (90%), followed by 'Mexico' (2%). There are 42 unique native countries.
• threshold: The binary target variable for income prediction, indicating whether income is 'over' (1) or 'under' (0) the specified threshold. 27,866 individuals are 'under threshold' (0), and 8,765 are 'over threshold' (1).

Distribution

Usage

• Developing and evaluating machine learning models for binary income classification.
• Conducting socioeconomic analysis to understand the influence of various demographic and employment attributes on income levels.
• Informing financial planning strategies by identifying factors associated with different income brackets.
• Supporting policy-making initiatives aimed at addressing income disparities or promoting economic welfare.

Coverage

License

Who Can Use It

• Data Scientists and Machine Learning Engineers: For building, training, and testing income prediction models.
• Social Scientists and Economists: To research and analyse socioeconomic factors influencing income distribution.
• Financial Planners: To understand income determinants for advising clients or developing financial strategies.
• Policy Makers: To formulate policies aimed at economic development, social welfare, or income equality.

Dataset Name Suggestions

• Adult Income Prediction Dataset
• Socioeconomic Income Classification
• Demographic Income Predictor
• Employment and Income Attributes Data

Attributes