After generating more than 1.5 billion pieces of content last year, Wordsmith’s prowess for generating human sounding narrative at scale is recognized around the world. While the sheer quantity and quality of the Wordsmith’s content is awe-inspiring, it’s Wordsmith’s unique ability to instantly process and manipulate countless rows and columns of complex data that makes us the most proud. Because at the end of the day, it’s the data that ultimately drives those billions of narratives.

So what do you need to know about your data sets before you get started?  To help, we’ve answered some of the most common questions we’ve received over the years.

And if you find that we didn’t answer all of your data questions, or you don’t have the resources to get your data into the right format, our Managed Services team can work with you to structure the data in a way that Wordsmith can utilize.

What are Wordsmith’s data requirements?

Acceptable Formats

Wordsmith requires your data set to be in both a structured and flat JSON or CSV format. What does that mean?

Structured Data 

Most commonly, structured data can be found in a spreadsheet (e.g. CSV or Excel file) or a relational database. It has a defined structure where specific values are categorized into distinct sets.

Flat Data

In addition to structured data, Wordsmith also requires that your data sets be flat. Flat data refers to data objects within a database or spreadsheet that have all corresponding data contained within a single row versus being nested or spread over multiple rows. Essentially, each row will be its own unique story and all of the data required for that story is contained in that single row.

Are There Size Limitations?

As a general rule, the more data is better. Since every word that Wordsmith generates is based on data, more data allows Wordsmith to compose with a greater sophistication. However, within a single project, Wordsmith limits the amount of key value pairs to 998 columns.

Does Wordsmith require historical data?


Historical data often improves Wordsmith’s output, but it is not required to for Wordsmith to tell a story. For example, our automated stories for the Associated Press compare current stock prices with historical prices:

United States Steel shares have decreased $1.83, or 6.2 percent, to $27.67 since the beginning of the year, while the Standard & Poor’s 500 index has increased 6.6 percent. However, the stock has risen $9.96, or 56 percent, in the last 12 months.

Can Wordsmith leverage multiple data sources?


If the multiple structured data sources can be aggregated into a single flattened source, that one source can be fed into Wordsmith to improve its insights. For example, Wordsmith can use real estate inventory data alongside data about local schools and restaurants to compose a richer narrative for potential homebuyers.

Want to learn more about how maximize success with Wordsmith? Request a demo to get a personalized review. Want us to handle everything? Head over to the Managed Services page to learn more.