Introduction

This repository contains the code for the paper "Estimating Global Patterns in Learning Quality Using Global Search Trends" by Serhat Arslan, Mo Tiwari, and Chris Piech, which appeared in Learning at Scale (L@S) 2020.

If you use code from this repository, please cite the following paper:

@inproceedings{arslan2020cslearning,
  title={Estimating Global Patterns in Learning Quality Using Global Search Trends},
  author={Arslan, Serhat and Tiwari, Mo and Piech, Chris},
  year={2020},
  publisher={Learning at Scale},
}

Requirements

The code requires Python 3.7.4 or above. The required python packages can be installed via pip install -r requirements.txt.

This project also requires chromedriver, specifically for Chrome v.75 on MacOS. Once installed, chromedriver must be added to your environment's PATH variable.

For further instructions and troubleshooting, please see: https://saucelabs.com/resources/articles/getting-started-with-webdriver-in-python-on-osx

Explanation of Files

• requirements.txt contains the necessary python dependencies; install with pip install -r requirements.txt
• constants.py contains the parameters for querying Google trends, including the string URL for each query
• downloader.py contains the code to loop through each search query on Google Trends and automatically download the statistics used in the computation of CSLI-s scores
• GT_Analysis_All_Final.xlsx contains the statistics downloaded from Google Trends using downloader.py, which are used in the computation of CLSI-s scores
• CLSI_scores_2014_to_2018.csv contains the computed CLSI-s scores
• cluster_data.py contains the code to plot countries' embeddings, both via an arbitrary 2D embedding and via t-SNE (typically applied to the CSLI scores)