OpenSpending is a flexible platform for storing government financial data. While we focus on collecting budget and transaction-level spending data, other types of financial reporting, such as accounts or cash flow information, can also be represented. To accomodate such different types of information, OpenSpending does not prescribe a specific format for the data that will be reported. Instead, we enable users to create a custom model for their dataset.
A typical workflow for importing a dataset into OpenSpending will therefore involve the following steps:
- Gather machine-readable data from a trustworthy source, collecting metadata such as the source, completeness and time covered in the dataset.
- Convert the data to a denormalized CSV file in which dates and numbers will be understood by OpenSpending.
- Upload the data to the web, e.g. to datahub.io.
- Create a dataset on OpenSpending and add the address of the dataset as a new data source.
- Model the dataset in OpenSpending to assign a logical role to each column in the source table. This includes designating the fields in which time, amount and, optionally, the spender and recipient of a transaction are specified.
- Load the data or refine the data conversion based on the feedback given by the platform about data consistency.
After the data has successfully loaded into OpenSpending, you can easily create visualizations, run searches or build custom applications on top of the APIs provided by the site.
TODO: Explain
- Facets for search
- Dataset naming convention
- Distinguishes capex/current
- Geocoding
- Maps as breakdowns