Help 4 Covid — A Step Towards Verified Covid Resources
We have all seen a bizarre time, thanks to COVID-19, over the last one and a half years. With the second wave of the virus hitting us like a storm, India had seen one of the worst phases of all time. It took a toll on all of us, both physically and emotionally.
With thousands of people posting everyday asking for help and only a handful were able to help with correct sources, we were upset on seeing that there was no common platform that everyone was aware of and where verified information could be updated.
The second wave resulted in a severe scarcity of resources. It was a mess, due to which many lost their lives. On the other hand, some good Samaritans manually verified resources (from vendors for oxygen to medicines and hospital beds in different cities). They updated this information on Twitter with relevant hashtags for others to find in case of any requirement.
Gagan and I realized that we wanted to make a change, a difference that would actually have an impact in this dreadful time. We were in two different cities at that time. He is a product manager in healthcare and life sciences, and I am a software developer.
After a couple of phone/zoom calls, we decided to develop something user-friendly and responsive. Our aim was to build a central point that connects people with leads to various resources (through Twitter) and has been more recently diversified into helping you find a vaccine slot for the pin code provided.
Our first version of Help 4 Covid needed to be up and running immediately to help the maximum number of people. It displayed all relevant tweets offering verified sources for oxygen cans/cylinders, amphotericin, hospital beds, ventilator units, ambulance, or food and then categorized them based on our pre-determined cities.
In the second version of our platform, we added the capability of finding vaccine slots based on your preference of vaccine and the pin codes you want to find in.
How does it work?
- We take the city and the requirement as input from the user.
2. We used the Twython API to fetch the relevant and recent tweets based on the user input. Along with the city and the requirement, we added the keyword ‘available’ and removed the retweets and tweets with keywords like ‘required’ or ‘requirement’ to narrow down the genuinely verified information.
3. We parsed the tweets to extract important information like timestamp, username, his/her screen name and the tweet text.
4. We used the web module in python to integrate our responsive UI and the python script.
5. Lastly, we hosted the website on pythonanywhere.com for free. It lets you track the usage and the logs (both server and error logs) for debugging purpose.
Covid vaccine slots availability
- User needs to fill the details like pin code, age and vaccine preference.
- The script used the requests module in python to fetch the data from Cowin APIs as provided on APISetu.
- As a website in pythonanywhere is hosted at a location outside India and there is a limitation of Cowin APIs that only IPs associated with India will be able to access them, we used the AWS EC2 instance to reroute the request so EC2 instance from Mumbai region hits the Cowin website to get the results.
- The IP address of EC2 instance was further added as proxies in the request that was sent.
Future Extensions
- Presently we only cater to 4 cities — Delhi, Bengaluru, Mumbai, and Kolkata, and the specific (most sought after) keywords. We intend to include more cities and requirements with NLP to filter out the most relevant tweets.
- Currently, only 100 tweets (most recent) are pulled up since there is a rate limit on Twitter APIs. In the future, we hope to optimize by storing the data into a DB and fetching them from the database along with a cron job in the background that can run in an interval to fetch new tweets.
Please feel free to leave any suggestion or feedback.
