With COVID-19 ravaging the world, we wanted companies reactions to be transparent. When a company does something to help the people in trying times, they deserve to be rewarded. So, we made Remembering COVID.
What it does
Remembering COVID aggregates, in real time, articles written about certain companies on various topics, such as covid response (covid, epidemic, pandemic), social altruism(donations, charity), and employer accountability(cancelling internships, forcing work from home). We assign scores using sentiment analysis and advanced numerical processing to each company on each of these three points and then average these scores for a final Remembering COVID score.
How I built it
We scraped the web using puppeteer/duckduckgo/chromium to get articles, and requests/BeautifulSoup to parse the body of these articles. We cleaned and joined the data with Python and Pandas. We used Google Cloud's Natural Language Processing API to perform the sentiment analysis and a Google Cloud hosted MongoDB atlas cluster to store all of the data. We wrote and ran all of the machine learning and data science code in a Google Cloud hosted Google Colab Jupyter Notebook. The frontend was then made in Vuejs with custom CSS styling with a tailored wireframe designed in Figma.
Challenges I ran into
Getting the websites to let me scrape them. Google Cloud doesn't like working when we need it most. I haven't slept at all tonight.
Accomplishments that I'm proud of
Fully finishing a project for once.
What I learned
We learned tools like Google Cloud, MongoDB Atlas, and UI Path.
What's next for Remembering COVID
Pushing this out to production and making it usable by the general public for years to come.
Try It out
beautiful-soup, figma, google-cloud, google-sentiment-analysis, mongodb-atlas, natural-language-processing, node.js, pandas, puppeteer, python, requests, vue