The Gateway to Algorithmic and Automated Trading

Quantopian's Fetcher applies any dataset to trading algos

First Published 2nd April 2013

Combination of any time-series data with price data opens new possibilities for quantitative trading

John Fawcett, founder and CEO, Quantopian

John Fawcett, founder and CEO, Quantopian

"The volume and diversity of the data available out there is unbounded and we're expecting some stunning results as people bake it into their algos."

Boston - Quantopian, which is building the world's first browser-based algorithmic trading platform, has announced the release of Fetcher, a new capability which will enable users to import any external dataset into the platform. With Fetcher, quants can combine data in new and innovative ways to reveal market insights and, ultimately, profits.

Fetcher allows any time-series, comma-delimited data to be pulled into Quantopian for analysis and insights. Whether it is first-party data users have themselves (an obscure index, commodity price, or derived trading signal for example) or third-party data sourced from the Web or elsewhere (demographic, pricing, meteorological, etc.) Fetcher can bring it into the Quantopian platform for use in algorithm development and testing.

"Our platform was a powerful tool when it used price data alone," said John Fawcett, founder and CEO of Quantopian. "With Fetcher, we've essentially turbo-charged the platform with the power to explore and execute on the furthest reaches of quant creativity. The volume and diversity of the data available out there is unbounded and we're expecting some stunning results as people bake it into their algos."

"It's exciting to consider the possibilities of the types of data that can now be easily combined into new alpha factors and backtested at the push of a button using Quantopian," said Jessica Stauth, director of quant product strategy at Thomson Reuters. "You can imagine investigating things like how the unemployment rate of young professionals affects the profitability of hotel chains. This opens some very interesting new areas for algorithm development and makes it even easier to go from idea to trade."

With so many openly available data sources such as Yahoo! Finance and Quandl - a searchable database of more than 4 million datasets - Fetcher opens algorithmic trading up to the full possibilities of creative financial minds. For example, Quantopian users can now import a record of the spot price of natural gas accumulated over the years to reveal how that price drives utility company stocks, or analyze how the historical trading price of gold correlates to stock performance of mining companies.

Fetcher places any imported time series in the dataframe with existing Quantopian-provided trade data. This ensures the same protection against look-ahead bias that is expected from a quality backtester. The introduction of Fetcher is another step in Quantopian's efforts to democratize finance by giving people access to the tools, capabilities and community they need to create and optimize their own trading algorithms in an open and transparent environment.