The code can be run either by running
python main.py or by executing the cells in main.ipynb after running
jupyter notebook main.ipynbFirst, create a virtual environment 'venv'. Then install the libraries by running:
pip install requirements.txtThe first time you run the program, you should uncomment the nltk package download code in main.py (you will also find it in the notebook if you prefer to run it via jupyter), then you can comment it out again.
There are two main configuration files: data/api_data.json and data/cryptos_to_analyse.json. The first one contains the set up API account for reddit, it is required in order to access their API. The second one contains a json-like representation of the three coins that will be analysed. For both files, I have provided some default values: for the api_data.json, I have already created an account and an API instance and provided its data, so you can directly use it. If you want - you can create a reddit account and go to the following address: https://ssl.reddit.com/prefs/apps/ and create another app that you hold, but it really doesn't matter. For the cryptos_to_analyse.json, I have provided three cryptocurrencies: bitcoin, ethereum and cardano. You can always change them if you want analysis for different ones.
Most of the results are represented visually, and all the relevant explanations can be found in the report Nexo_crypto_task_report.pdf. The report is written in LaTeX, using OverLeaf. In short, the code produces some graphs in the visiualization folder. The graphs are mainly about sentiment, word frequency, and word count.