tw-insights

Analyze a tweet archive using Microsoft Azure's text analytics services.

Getting data

You will need to get an archive tweets to analyze. The easiest way to do that is to get your own archive from can do that at https://twitter.com/settings/account

Microsoft Azure credentials

You will need to create a Microsoft Azure account, enable Cognitive Services and then place your API key in a .env file like so:

MS_ACCESS_KEY=access-key-value

More instructions can be found here: https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/how-tos/text-analytics-how-to-signup

How to run

There are two ways to run this tool

Top-level command

$ tw-insights run-all [path-to-archive]

This will read all the tweets in the archive and run them through all the analysis steps. In doing so, it will split the tweets into batches of 1000 to respect Azure's upload limit. The tool does nothing else to manage rate limiting, so you might hit errors if you analyze a lot of data at once.

Individually run steps

$ tw-insights read-tweets [path-to-archive] | tw-insights add-languages | tw-insights add-sentiment | tw-insights add-key-phrases

The individual commands have all been built to read tweets from stdin as new line delimited JSON. The results of one command can be pipe into another, and the tool will automatically batch them into groups of 1000 per request to respect Azure's upload limits. This approach may lead to better performance on large datasets. Note that sentiment and key phrases analyses require a language to be set, which language analysis step provides.