Scripts to build dataset of users from both Instagram and Twitter

1. Python & PHP (we used Python 2.7 and PHP 5.6)

2. instagram-php-scraper

   It is already required in the composer.json file, so all you have to do is

   composer install
   

3. python-twitter

   To install using pip:

   pip install python-twitter
   

4. FacePlusPlus Python SDK

   Place the facepp.py in the same folder

1. Go to http://apps.twitter.com and create an app to obtain consumer key, consumer secret, access token, and access_token_secret. Fill these four values in search.py
2. Run the following lines in a sequential order in your Terminal or Command Prompt

python search.py
python aggregate_users.py
python detect_age.py
php getMediaByUrl.php
python assemble_dataset.py

Each line can be stopped when you already have enough processed data. The output(s) that are important are:

• users_instagram.json: raw data in JSON format
• users.csv: processed data in CSV (only contains features which are mentioned in the paper)

Because the dataset contains privacy sensitive information of users, if you would like to request the full dataset file, please send an email to e.a.rizkiasri@student.tudelft.nl stating your purpose of using the dataset.