If you’re the TA correcting our work, you can directly skip to the second part describing our approach: chances are you already know what the project is about :)

The project

We are given a dataset and asked to come up with a project from it. Within a semester, our task was to get acquainted with how the dataset was constructed, what was inside it, how the data was gathered, and what we could do with it. First, we came up with bold and creative ideas of projects that could be done using this dataset. Then, we formed the team, selected the project we wanted to do and performed initial analyses of the data. The end of the semester was dedicated to the concrete realisation of the project, and the creation of the datastory, which can be found on this website.

Our approach

The beginning of the project was a bit vague since there were no clear instructions of what we should be doing. Our ideas of projects received positive feedback at first, even though it appeared we were a little optimistic about what could be achieved in the allocated time. Then, we might have gotten too confident, and planned on analyzing a lot of human behavior : our goal was to assess the evolution in time of the perception of events by social groups, through case studies. Ultimately, we hoped to maybe uncover hidden patterns in human behavior. The feedback we got incited us to be a little less ambitious since human emotions are one of the hardest things there is to model. In the end, we decided to analyze the human perception of two car manufacturers, with very different profiles. In essence, we wanted to compare electric cars and thermic cars. To perform this analysis, we used several techniques on specific subsets of the dataset, such as Sentiment Analysis (for which we used three different tools), and Topic Analysis, which allowed us to detect events only by scanning and analyzing quotes.