This is an online tool that can be used to reformat, clean up and organise your dataset.

Once you have gathered your data and entered it into a spreadsheet you are left with the laborious, yet essential, task of cleaning it up. A messy dataset does not only hinder you from undertaking a meaningful analysis, but it can also prevent you from being able to use visualisation tools or statistical packages as many require specific formatting. Most datasets need some sort of 'cleaning up' to improve their quality; issues such as misspellings, missing data and duplicated entries need be identified and fixed.

Rather than the traditional method of manual-editing, Data Wrangler offers 14 different options for formatting your dataset and within each of these options are a variety of choices. These functions let you split, merge, delete, promote, wrap, fold and unfold your rows or columns and also to cut, extract and edit your data. This variety of choices may seem overwhelming, but the tool helpfully suggests possible changes in rank-order and also offers visual previews so you can try the different options out before accepting them.

Data Wrangler enables you to reformat your data values or the layout of the dataset, amend incorrect or missing values and split entries into separate columns or rows. For example, if you were to choose the 'unfold' option it would allow you to identify different value types in one column and separate them out into new columns. Another example is the 'fill' option that identifies empty cells in your dataset and copies and pastes information contained in the preceding non-empty cells into these empty cells. This is a helpful function when dealing with visualisation tools as many will not work if a dataset contains empty cells, as they do not recognise them.
It is often difficult to discover data quality issues and usefully Data Wrangler highlights errors in red and indicates whether a column has inconsistent data types and flags up the individual potential errors.

The tool has a interactive history function which keeps a record of the changes performed and also allows you to annotate the rationale behind the changes. This is useful when multiple analysts are using the same dataset and want to review, reuse or modify the previously applied changes to suit their particular needs and interests. An aspect of this tool worth highlighting, especially for those working with large datasets, is that you can choose to export your changes as code through either Python or JavaScript. Data Wrangler has a limit to the amount of data you can reformat using their tool, but this feature enables you to edit a sample of your data in Data Wrangler and then export the resulting script to use offline on your full dataset.