Dow Jones Equips Developers with Sample Codes

​Dow Jones is publishing developer blueprints—pre-packaged source codes—to help software developers and data scientists reduce the amount of time it takes to perform high-quality API documentation and deployment by using sample codes.

More broadly, the blueprints are meant to help developers improve the quality of operational and strategic decisions made in their firms.

Niranjan Thomas, general manager of Dow Jones’ developer platform and solution engineering team, says the company felt that developers needed more contextually relevant tools to solve their engineering problems. Thus, the team began brainstorming blueprints, also called “solution patterns” internally.

“In other words, [they are] business domain-agnostic solutions made for developers to re-purpose and re-use. Additionally, it allows us to help developers with different levels of expertise: An expert developer and architect could just read our solution pattern write-up, diagram, and adapt to their internal stack; a more junior developer might need the sample code to get started faster,” Thomas (pictured) says. 

He says the engineering team is in ongoing talks with the global fintech, marketing tech, and regtech developer communities to identify ways that Dow Jones’ products and developer platform can improve productivity. 

Dow Jones publishes solution patterns under three main categories: data science, data management, and user-interface design. The blueprints leverage open-source libraries and tools so that developers can—and are encouraged to—use their internal data and other third-party datasets, on top of those provided by Dow Jones.

“Our datasets are really a ‘multiplier’ of the datasets developers or data scientists are already using,” Thomas says.

The Covid Blueprint

One of the latest blueprints Dow Jones launched is the Covid-19 Timelapse dashboard, which tracks the coronavirus’ impact on individuals, industries, and regions around the world, leveraging Dow Jones data alongside public health statistics. In conjunction with proprietary news and industry information, Timelapse runs on the same public datasets⁠ used to power the coronavirus-tracking dashboard operated by Johns Hopkins University, as well as data compiled by the Harvard Global Health Institute.

It provides a daily view of criteria such as virus spread by region, strain on health systems, country-based news coverage, new virus-related terms and definitions, and news sentiment by industry.

“The dashboard itself provides what we call a dynamic time-lapse view,” he says. “Since its launch, we are seeing users really focus on exceptions that are affecting industries of interest. For example, that could mean something like positive or negative sentiment changes across multiple industries that have been impacted by the pandemic. The dashboard also highlights contextually relevant news stories for users who wish to drill down further into trusted sources from Dow Jones or other publishers.” 

The Covid-19 Timelapse dashboard is categorized as a “user-interface design” solution pattern, and the sample code is available for download at Dow Jones’ developer platform GitHub repository. Thomas says the intention is to equip developers with tools needed to build their own unique dashboards based on questions their companies are trying to answer, but software developers, data scientists, and UX/UI designers from any industry can use the collection of data to find patterns and correlations, and cross-reference one collection with another to generate comparisons. 

“For example, [users are] comparing topic coverage between a specialized industry publication and mainstream published content. This broad dataset is incredibly important for understanding correlation, causation, and network effect of events that impact capital markets,” he says. 

The dashboard uses Plotly’s Dash framework and Mapbox for visualization and the user interface. For data preparation and engineering, it uses Python scripts that run in the Flask web app framework in the Heroku container. 

Coupled with Dow Jones’ Factiva Snapshot API, developers get a moment-in-time snapshot of content from over 8,500 data sources within the Factiva database. The API is built on Google Cloud, which allows a Snapshot extraction with hundreds of millions of news articles to be carried out in a few minutes. For sentiment and topic analyses, it uses open-source natural language-processing tools, including TextBlob.

“All of these elements can be swapped out with more powerful capabilities that a developer may be more familiar with,” Thomas says. 

Interested users can access the blueprints at the web domain: developer.dowjones.com.