Palantir

Metropolis

Integrate, enrich, model, and analyze any kind of quantitative data.

Applications

On the front-end, Palantir Metropolis provides a suite of integrated applications that empower users to build up interactive models using all the data that's available in their system, a rich library of metrics, and a programming environment to create novel metrics.

Custom Metric

The Custom Metric application includes a rich editing environment that allows analysts to extend the base functionality of the platform by creating new metrics to perform novel, proprietary, or specialized analysis. From something as simple as a spreadsheet formula to a more sophisticated block of code, Custom Metrics can be used to rapidly build up a specialized analytic vocabulary that makes it easier to conduct complex analysis.

Dashboard

The Dashboard application allows users to visualize collections of analysis on a single screen. The Dashboard acts as both a place to consume the output of complex analysis and a creative springboard for new lines of inquiry. Widgets displaying a mix of visual and textual data are arranged in a customizable layout. Each widget is interactive—users can adjust dashboard views in an ad-hoc way, or jump directly into source documents to iterate on analysis. Views update in real time as new data arrives into the platform. Users can also publish read-only, web-accessible versions of their dashboards to easily share finished work across an organization.

Date Set

When analyzing quantitative data, it is useful to have "time" as a first-class object. The Date Set application enables analysts to create regions of time that can be used as inputs to metrics and documents. Time regions can be chosen using periodic patterns like, "the third Tuesday of every-other month," or, "every ten seconds." More powerfully, time regions can be specified using computed metrics.

Some notional examples are:

  • All times when series A is greater than series B
  • All times when series X crosses its +2 sigma bollinger band using a 90-day window
  • All times when the aggregate value of the top ten series in set _T&Q 400_ is more than ten times the aggregate value of the bottom ten series and the aggregate value of the set is increasing

Explorer

The Explorer application is used to drill down on and save an interesting set of models for analysis. Explorer lets the analyst define and apply a series of filters over a large set of models. The resultant sub-set of models can be saved as a group and used in other applications or in HHLang expressions. Users can filter on model metadata (e.g., the type of model, data source, etc.) or apply parametric filters using computer metrics. Analysts can leverage interactive visualizations like histograms, scatterplots, and quantile charts that make nuanced selection quick and accurate.

Regression

The Regression application provides an interactive environment where analysts can build up regression models to run against any data in the system. Each aspect of the regression model is visualized as various inputs and parameters are adjusted. Documents produced in the Regression application expose various residuals and model tests as elements that can be used in further analysis.

Spreadsheet

The Spreadsheet application can incorporate and visualize models, groups of models, and Palantir Metropolis documents in its cells. Source documents for each cell's data are just a click away, making spreadsheets a way to centrally collect and compare the building blocks of a complex piece of analysis.