A Twitter live-data collector. Keep an eye on your disk space ;-)

This project is largely an experiment that may one day grow into something useful for more than just a handful of people.

The idea is simple: Albatross is a way for you to capture everything said about something and then analyse the raw data yourself or let this site visualise it for you.

Say you're at a conference that's managing the conversation with the hashtag: #awesomeconference. You're participating when you can, but it would be really nice to be able to collect everything everyone said during the conference and draw some conclusions from it.

Maybe you want to do some analytics based on some natural language processing, or want to chart the number of times a particular phrase was mentioned within that hashtag. Whatever you want to do with the data, just fill out the form here with the hashtag in question, hit submit and when the conference ends, you've got all the data to play with.

Similarly this can be used for international events, or national disaster coverage. Plot on a map the tweets posted about your subject and when, or analyse the content of the tweets to see what different regions are saying about a particular subject.

Analytics and visualisation is hard, and not everyone has the time or inclination to figure out how to parse the JSON blobs Twitter makes available via their API. Thankfully, Albatross has a bunch of built-in visualisations that might be sufficient for many people, but if you need something more customised, you can open an issue and I can give it a shot. Of course, contributions are welcome!

The raw data is available as a xzip-compressed "fjson" file. This is just a plain text file, with one JSON object per line. You can decompress this file in Linux & Mac with the xz utility, or use a common program like WinRar in Windows.

Everything is dockerised, so you need:

• Docker

• Docker Compose

• A .env file defining the following values:

  PYTHONUNBUFFERED=1
  
  SECRET_KEY=<secret>
  DEBUG=false
  ADMINS=[["Your Name", "your@email.address"]]
  DJANGO_LOG_LEVEL="INFO"
  
  TWITTER_CONSUMER_KEY=<secret>
  TWITTER_CONSUMER_SECRET=<secret>
  

  Note that the values aren't wrapped in quotes. You can thank Docker for that.

  Also note that you can configure the externally-available prots for the web service and RabbitMQ by setting WEB_PORT=<number> and RABBITMQ_PORT=<number> respectively.

$ git clone https://github.com/danielquinn/albatross.git

$ cd albatross
$ vim .env

$ docker-compose up

Docker makes this easy:

docker-compose up

When it's finished, you should be able to visit http://localhost:8000/ and then login with Twitter to start your first collection.

Here's what it's doing under the hood:

I wrote this back in 2015 as an attempt to start a company doing this sort of thing for people & businesses. However, not long after I had a working model, Twitter started acting more and more hostile to projects that would dare store "their" data, so I gave up on it.

In 2018 however, I repurposed Albatross into a self-hosted model to help a friend do some research for her PhD. If it works for her, it may work for others, so I decided to polish it up a bit and re-license it under the AGPL3.