This package, when run, will execute a query every day that collects all tweets from Wales for the past 7 days (if running for the first time), or since the last tweet collected. It uses Twitter's V2 API.
When run it will set up an SQLite3 database in the location specified to the main script which saves tweets, their basic information and their sentiment scores from the VADER sentiment algorithm in a table called tweets.
A table called places then saves the geo expansion information requested in the query.
Finally, a table called matchedplaces saves the result of each place matched to a Welsh local authority using an algorithm defined in the tweets/geo.py file.
This package was written by ninadicara, altanner, and leriomaggio.
This uses the academic API, so you need an approved account with a bearer token.
If you'd like this to work out the box then you'll need to set
the bearer token as an environment variable called SEARCHTWEETS_BEARER_TOKEN.
Currently the query returns tweets with basic information and requests the geo expansion.
If you change the query you'll also need to edit the SQL tables and data entry functions.
These are in tweets/database.py, called create_tweets_tables() and add_tweet_json().
The database is SQLite3, for ease and as an alternative to CSV.
The schema for the database can be seen in the setup functions as part of the Database class in tweets/database.py.
There is a foreign key between place_id in the tweets and matchedplaces tables that link to the primary id entry in the places table.
Source: Office for National Statistics licensed under the Open Government Licence v.3.0.
Contains OS data © Crown copyright and database right 2019.
This package uses boundary data of Welsh Local Authorities (see utils/la_keys.geojson).