generalmills/pytrends

Pytrends is an unofficial Python library that lets you download data from Google Trends without having to click through the website manually. Google Trends shows how often people search for particular words or topics over time, broken down by region, time period, and search type (web, news, YouTube, images, or Google Shopping). Pytrends gives you that same data in a format you can work with in Python, returned as tables (pandas DataFrames) you can analyze, chart, or save to a file. You install it with pip, create a connection object, tell it which keywords you want to track (up to five at a time), and then call whichever data method you need. The available methods include interest over time (a historical index of how search volume changed), interest by region (which countries or states search for the term most), related topics, related queries, trending searches, top charts, and keyword suggestions. There is also a method for pulling hourly historical data, which requires multiple requests to Google since Trends only exposes one week of hourly data at a time. The library is unofficial, meaning it works by mimicking the requests a browser would send to Google Trends rather than using a supported API. Google can and does change its backend, which can break the library without warning. The README acknowledges this directly and invites contributors to help fix it when that happens. Proxy support is built in for situations where Google rate-limits your IP address. Search terms can be either plain text or Google's internal topic IDs. Using topic IDs is more precise: searching for "iron" as plain text will match everything called iron, but the topic ID for "Iron (chemical element)" targets only that concept. The library includes a suggestions method to help you find the right topic ID. The project lists General Mills as the maintainer on GitHub, though it appears to be community-driven. As of the README, the project is looking for new maintainers.