The Good Tables web service

Posted on 06 March 2015 by Paul Walsh

Introducing the Good Tables web service

Good Tables is a free online service that helps you find out if your tabular data is actually good to use - it can check for structural problems (blank rows and columns) as well as ensure that data fits a specific schema.

Tabular data in CSV and Excel formats is one the most common forms of data available on the web - especially if looking at open data. Unfortunately, much of that data is messy with blank and incorrect rows, and unexpected values for some fields. (For example, date columns that do not feature well-formed dates. See here for more examples of “bad data”.)

That’s where Good Tables comes in: it checks your data for you, giving you quick and simple feedback on where your tabular data may not yet be quite perfect.

Good Tables uses the previously announced Good Tables Python library, and is developed by Open Knowledge with funding from the Open Data User Group.

Good Tables is currently an alpha release; we invite the community to start using and contributing to it to help us move towards v1.0.

API

The documentation for the API can be found here.

Using the API is easy: POST or GET your data, and get back a JSON object containing the report.

For example:

# make a request
curl http://goodtables.okfnlabs.org/api/run --data "data=https://raw.githubusercontent.com/okfn/goodtables/master/examples/row_limit_structure.csv&schema=https://raw.githubusercontent.com/okfn/goodtables/master/examples/test_schema.json"

# the response will be like
{
    "report": {
        "summary": {
            "bad_row_count": 1,
            "total_row_count": 10,
            ...
        },
        "results": [
            {
            "result_id": "structure_001", # the ID of this result type
            "result_level": "error", # the severity of this result type (info/warning/error)
            "result_message": "Row 1 is defective: there are more cells than headers", # a message that describes the result
            "result_name": "Defective Row", # a human-readable title for this result
            "result_context": ['38', 'John', '', ''], # the row values from which this result triggered
            "row_index": 1, # the idnex of the row
            "row_name": "", # If the row has an id field, this is displayed, otherwise empty
            "column_index": 4, # the index of the column
            "column_name": "" # the name of the column (the header), if applicable
            },
            ...
        ]
    }
}

For more details on the report response, see the report section of the Good Tables documentation.

UI

The web service also features a form for manual validation of data via a UI.

Let’s see it in action:

Contributions

We invite all contributions. Feel free to open an issue if you encounter any problems, or just start hacking and send a pull request.