Quantopian is a cloud-based platform that allows you to research and test trading algorithms using all sorts of datasets. When researching a strategy on Quantopian, the majority of your time will be spent in Research and the IDE. However, there are certain circumstances where you might want to do some of your work in a local environment (your own computer). This section outlines the differences between working on the web platform and working locally, and enumerates some of the situations where local development might be preferred.

Online vs. Local

There are a couple of major differences between working on the Quantopian platform and working locally:

  • Data: Pre-integrated datasets are easily and freely available on the platform, but cannot be downloaded locally. To work locally, you need to provide and clean your own data.
  • Tools: All of Quantopian's tools and whitelisted modules come pre-installed on the platform. To work locally, you'd need to install the packages yourself, which often requires prior experience with Python, bash, and package management. And some tools (like Optimize) are proprietary, so they can't be installed off of the platform.

For these two reasons, we generally recommend that you use the Quantopian platform for most of your workflow.

However, there exist some cases where it might be appropriate to complete some steps of your algorithm development workflow in your local environment. The next section describes why you would need to work offline, and how you can set up your offline workflow if needed.

Reasons to Work Locally

Currently, creating custom datasets for Self-Serve Data is the primary use case for local development. It's useful to work locally when creating a dataset for Self-Serve for a couple of reasons:

  • Self-Serve Data requires you to upload a .csv file from your local machine. Since downloading data from Quantopian is not supported, you will need to collect your own data and create the .csv file locally.
  • Dataset construction often needs modules that aren't allowed on Quantopian (for example, the requests library for web scraping).

Local Installation

If you do need to work locally, begin by determining which packages you need. Then, install the packages.

Available Quantopian Tools


This section lists open-source Quantopian tools. You might also want look into installing pandas, statsmodels, numpy, Jupyter and other data/statistics packages.

The following Quantopian tools are open-source and available for local installation:

The following tools are proprietary and not available for local installation:

  • Optimize
  • Risk Model
  • Research API

(While the Research API is not proprietary, it depends on an extensive backend that cannot be setup locally. For a local alternative, try Jupyter.)


To install, simply follow the installation instructions for each package you need.

Helpful links: