The DC Purchase Card Transactions Dataset is provided by the DC Government as part of an effort to promote transparency and accountability. The dataset includes over 500,000 transactions and is updated monthly by the Purchase Card Program Management Office, part of the Office of Contracting and Procurement.
The dataset includes the following key fields: Agency, Transaction Date, Transaction Amount, Vendor Name, Vendor State/Province, Merchant Category Classification Description, and Last Modified Date. The data has been imported into a Neo4j graph database for efficient querying and visualization.
Currently, the following three tools are available:
Interactive Data Explorer
Users can freely query the data and visualize results as an interactive tree graph. Clicking on a node dynamically generates a time series, histogram, and key statistical metrics in real-time. Additionally, users can download the corresponding data in JSON and CSV formats.
Local Outlier Factor (LOF)
The LOF tool identifies anomalies in the transaction data of a specific agency. It calculates an LOF score for each transaction, highlighting those that deviate significantly from normal patterns. Transactions with high LOF scores may indicate unusual activity, such as exceptionally high amounts or infrequent vendors.
Isolation Forest Tool
The Isolation Forest (IF) tool detects anomalies in transaction data by calculating an anomaly score for each transaction. Transactions with lower IF scores are flagged as potential outliers, suggesting unusual activity such as exceptionally high amounts or infrequent vendors.
Use Cases
Explore various ways to utilize these tools in the Use Case section.
For more details on the DC Purchase Card Transactions, visit the official DC Open Data Portal.
DISCLAIMER: The tools and the corresponding database are based on the original DC Open Data. However, the accuracy and completeness of the data generated by the tools cannot be guaranteed.