Sentiment Analysis of Tweets in Tunisia: Evaluating the Effect of an ENPI Project using R

In this paper, I examine empirically the impact of an ENP human-rights project, by analyzing French and Arabic tweets of people from three towns in Tunisia where the project was implemented and comparing their content with the tweets of people from three towns where the project was not implemented. This research initiative constitutes my first attempt at sentiment analysis using software – I elected to use R as I have been familiar with it but have struggled with carrying out the Arabic text analysis. Despite that, the preliminary results of my study indicate that there is indeed a more pro-human rights discourse conducted in the towns where the project in question was implemented as opposed to the control towns, even though whether this reality can be attributed to the project is debatable. I want to continue with the analysis to find further ways of controlling for confounding effects.