askcomm: Python 3 module - Search patterns for event-based, community-detected twitter data.

By Chris Lindgren chris.a.lindgren@gmail.com

Distributed under the BSD 3-clause license. See LICENSE.txt or http://opensource.org/licenses/BSD-3-Clause for details.

Overview

A set of search patterns that query a corpus of event-based and community-detected tweets, but it could be modified to query most social-network (node-edge) data. The queries are great for content produced within the detected-community subgraph data.

It assumes you have:

• imported your corpus as a pandas DataFrame,
• included metadata information, such as a list of dates and list of groups to reorganize your corpus, and
• pre-processed your documents as community-detected data across periodic events.

Functions

query_controller: Accepts corpus and hub user data and searches for tweets germane to the detected module community across a range of periods and communities. It uses the find_mentions function to conduct a cross-reference search within a period's data range with 2 options: 'mentions_only' or 'user_and_mentions'. 'mentions_only' searches a column with a List of mentions per tweet. 'user_and_mentions' cross references the author of a tweet with the list of mentions. It returns a Dict of top result tweets found during that period.

query_controller(
    hubs=df_hubs,#community-detected data
    hub_col_period='period',#column name for periods
    hub_col_module='info_module',# column name for community name
    hub_col_users='name',#column name for 
    period_range=[1,10],#range of desired periods
    module_range=[1,10],#range of desired communities/modules
    corpus=c_htg,#content corpus
    period_dates=period_dates,#List of lists with dates to 
    col_dates='dates'#column name for dates
)

convert_to_df: Converts the Dict output from query_controller into a Dataframe with top result per user. If no tweet found , appends as None.

find_ht: Queries subset of isolated mentioned or authored tweets with hashtag group list. It returns another subset as a dataframe.

find_links: Queries links in tweets with search string. It returns subset as a dataframe.

Other functions include: find_mentions and print_subset.

It functions only with Python 3.x and is not backwards-compatible.

Warning: askcomm performs little to no custom error-handling, so make sure your inputs are formatted properly. If you have questions, please let me know via email.

System requirements

• pandas

Installation

1. Download this repo onto your computer.
2. Store the folder in a meaningful location.
3. Open a terminal.
4. In the terminal, navigate to the root of the folder.
5. In the terminal, run pip install .

Known Issues or Limitations

• Please contact me if you discover any issues.

Example notebooks

• Coming soon.

Distribution update terminal commands

# Create new distribution of code for archiving
sudo python setup.py sdist bdist_wheel

# Distribute to Python Package Index
python -m twine upload --repository-url https://upload.pypi.org/legacy/ dist/*