This repository contains materials to reproduce and replicate an analysis on using geosocial sensors.
List of files and brief description:
TwitterStreamingAPI_London.py
Python script used to retrieve Tweets
FlickrSearchAPI_London.py
Python script used to retrieve Flickr image meta data
TwitterBuildTermVectors.py
Python script used to parse Tweets for terms found in terms_stemmed.txt
FlickrBuildTermVectors.py
Python script used to parse Tweets for terms found in terms_stemmed.txt
Flickr_IDs.zip
CSV file with IDs of all Flickr images used as input data
Tweet_IDs.zip
CSV file with IDs of all Tweets used as input data
bots_identification.sql
SQL statements used to determine which Tweets are likely to be from bots (and excluded from further analysis)
TF-IDF_computation.ipynb
Python Jupyter notebook which computes TF-IDF scores
TF-IDF_analysis.ipynb
Python Jupyter notebook which analyses the TF-IDF scores and socio-demographics
*.npy
several files containing intermediary results (numpy arrays of the TF-IDF scores)
Flickr*.csv and Twitter*.csv
several files containing results of the TF-IDF scores analysis
sentiment_conputation.ipynb
Python Jupyter notebook which computes sentiment scores
vaderSentiment_mod.py
modified VADER sentiment anlyzer (see cjhutto/vaderSentiment#99 and paper for details)
sentiment_analysis.ipynb
Python Jupyter notebook which analysis the sentiments and socio-demographics
sentiments_groups_wards.csv
results for sentiments and LOAC groups at ward level
sentiments_sociodem_msoa.csv
results for sentiments and socio-demographics at MSOA level
spatial_temporal_analysis.ipynb
Jupyter Python notebook which analysis sentiments for spatial pattern and changes over time
environment.yml
Miniconda environment file