Digital_Payments_2025_Dataset
Description
The "Digital Payments 2025 Dataset" is a synthetic dataset representing digital payment transactions across various payment applications in India for the year 2025. It captures monthly transaction data for multiple payment apps, including banks, UPI platforms, and mobile payment services, reflecting the growing adoption of digital payments in India. The dataset was created as part of a college project to simulate realistic transaction patterns for research, education, and analysis in data science, economics, and fintech studies. It includes metrics such as customer transaction counts and values, total transaction counts and values, and temporal data (month and year). The data is synthetic, generated using Python libraries to mimic real-world digital payment trends, and is suitable for academic research, teaching, and exploratory data analysis.
Files
Steps to reproduce
1. Install required Python libraries (pandas, numpy) by running the command pip install pandas numpy in a Python 3.8+ environment to handle data manipulation and random number generation. 2. Create a list of 80+ payment apps, including major UPI platforms like Google Pay, PhonePe, and Paytm Payments Bank App, as well as smaller bank apps like Allahabad Bank App and Janta Sahakari Bank App, to represent the diversity of digital payment services in India. 3. Define the time period as January to December 2025, represented as a list of months (1 to 12) and a fixed year (2025), to structure the dataset with monthly granularity. 4. Initialize an empty pandas DataFrame with columns PaymentApps, CustomerTxCount (Mn), CustomerTxValue (Cr), TotalTxCount (Mn), TotalTxValue (Cr), Period, and Yr to store the synthetic transaction data. 5. Generate synthetic transaction data for each app and month by iterating through the app list and months, using numpy.random.uniform to create realistic transaction counts (0.01–2000 million for customer transactions, higher for major apps like PhonePe) and calculating transaction values based on an average transaction size (₹100–₹10,000, varying by app type). 6. Adjust total transaction counts and values by applying a slight variation to customer counts and values using numpy.random.normal (mean 1, standard deviation 0.05) to simulate non-customer transactions, ensuring total counts are within ±5% of customer counts. 7. Apply realistic trends by scaling transaction counts and values for December by a factor of 1.2 to reflect increased activity during the festive season, using a conditional check for month 12. 8. Simulate incomplete data for smaller apps by randomly dropping 1–3 months of data for 10% of apps (e.g., Janta Sahakari Bank App) using numpy.random.choice, ensuring at least 50% of apps have data for all 12 months to maintain dataset robustness. 9. Save the generated DataFrame to a CSV file named "Digital_Payments_2025_Dataset.csv" using the pandas to_csv method with index=False, ensuring the output matches the provided dataset structure (approximately 900 rows, 7 columns). 10. Validate the generated dataset by checking that major apps (e.g., Google Pay) have significantly higher transaction counts (500–2000 Mn) and values (10,000–400,000 Cr) compared to smaller apps (0.01–0.1 Mn, 1–50 Cr), and confirm no critical fields are missing.
Institutions
- Bharati Vidyapeeth Deemed University College of Engineering Pune