Sleep Health Disorder Dataset

Dataset Overview The Sleep Health and Disorder Dataset (SHDD) is a survey-based dataset comprising 2,119 individual records collected to support research on sleep disorder classification. The dataset captures demographic, physiological, and lifestyle-related variables associated with sleep health, along with a diagnostic label indicating the presence and type of sleep disorder. Data Collection All records were collected independently through a structured survey covering demographic information, sleep patterns, physical activity, stress levels, and cardiovascular indicators. No existing public dataset was used or modified; the feature schema (variable categories) follows commonly used sleep-health research variables, while all data values were independently collected. Dataset Structure Total records: 2,119 Total features: 13 (12 predictor variables + 1 target label) Feature Description Gender Male / Female Age Age of respondent (27–59 years) Occupation Profession (10 categories, e.g., Nurse, Doctor, Engineer, Teacher) Sleep Duration Average hours of sleep per night Quality of Sleep Self-rated sleep quality (scale 1–10) Physical Activity Level Minutes of physical activity per day Stress Level Self-rated stress level (scale 1–10) BMI Category Normal / Overweight / Obese Heart Rate Resting heart rate (bpm) Daily Steps Average daily step count Systolic_BP Systolic blood pressure (mmHg) Diastolic_BP Diastolic blood pressure (mmHg) Sleep Disorder Target label: None, Insomnia, or Sleep Apnea Class Distribution None (no disorder): 968 (45.7%) Insomnia: 580 (27.4%) Sleep Apnea: 571 (27.0%) Important Note on the "Sleep Disorder" Column The target label "None" indicates the absence of a sleep disorder and is a valid class category — it is not a missing value. When loading this dataset with pandas, use pd.read_csv(filename, keep_default_na=False) to prevent the string "None" from being automatically interpreted as NaN. Data Quality No missing values No duplicate records No negative or physiologically implausible values All blood pressure readings satisfy Systolic > Diastolic Pulse pressure values fall within the normal physiological range (21–61 mmHg) Intended Use This dataset is suitable for multiclass classification research, exploratory data analysis, and machine learning model development related to sleep disorder prediction, as well as studies examining the relationship between lifestyle factors and sleep health outcomes.