In Real Life Science World

COVID-19 Predictive Analysis

How do you think the situation regarding the COVID-19 pandemic will change in the coming few months? When will it finally come to an end? When can we resume our normal lives? The answer to many such questions is contained in the link below! Make sure to click it and give it a read to find the answers!!

Click the dowload icon above to save and read my assessment on the current COVID-19 situation in India and what I think will happen in the near future or just click the link beside it to read it online.

Incase you want to see the working of the predictor, I have attached a link below as it might not open from the PDF in some devices.

https://growthratepredictor.aaryannagpal.repl.run
(Just click on the screen and enter answers to the questions the program asks to get projections for the future of COVID-19.)

The research asks and answers some really important questions-
1) How effective were the lockdowns?
2) When will the curve plateau?
3) Does the virus affect all people equally?

Alongside many such other questions, the research also provides a detailed analysis on how the number of cases will increase, when will they start to decrease, how high the recovery rate is and how does the death rate of India compare to other countries.

A detailed mathematical and statistical model compiled with the use of python provides me the oppurtunity to graph my projections for all such parameters.

Some frequently used websites for data inputs were-
https://www.covid19india.org/
https://www.worldometers.info/coronavirus/

Hope you have a wonderful time reading the research! Drop any suggestions in the comments or link a section from the research you might want a secondary article on. Share this post with as many people as you can!

91 comments

  1. This is an extremely detailed article. Nuances of pre/post lockdown are captured, and analytics on Indian data and hotspots is intriguing. Way to go Aryan.

    Liked by 1 person

  2. Good Job Aaryan. your Analysis is pretty good and the content is very good. But i think bit more details required to explain some abbreviations you have used will help the readers.
    See if you can use any other option than REPL for user interaction.

    Liked by 1 person

    1. Thank you! I tried to include all keywords, characters and parameters I could think of in the 2nd page, I will be looking for more to add. I will be trying to use interactive softwares with a better UI/UX. I’m working on that!

      Like

  3. I must say it’s one of the most extensive fact based research that I have gone through in last couple of months . It is written based on hardcore mathematical models and the use of technology had been icing on top of the cake. A real good job and a big thumbs up for the same . As a next step may be you can write on the implications of the same what are possible impacts in the directions of Economy / Life’s of People & lifestyle changes which will become the new norm etc.

    All in all a fantastic Job keep up the good work.

    Liked by 1 person

    1. Thank you for the review and the suggestion! I have a new post coming soon, will try to include some of the suggested topics in that.

      Like

  4. Good job in reasearch and formulation on number of cases and forecasts. Can you please also add actuals based on some real numbers as examples and see how it compares to your analysis based of formula?
    Also some information on vaccine progress by various institutions would be nice idea.

    Liked by 1 person

    1. Thank you for the suggestion! Figures 2,4 and 8 cover analysis of the actual numbers with those projected by the predictor. I will be including more comparisons in the new posts, the details on the vaccine will also follow in the coming posts!

      Like

  5. The article has been very nicely organized and articulated. Assumptions and deductions are also pretty logical. A first impression would suggest to use exponential distribution, but since we are using different values of k for different weeks, that gets covered with a linear variation. If possible, an analysis at the “day level” granularity would have even been more accurate. Really in-depth analysis. I wish you that this model shows very little deviations from actuals !!!

    Liked by 1 person

    1. Thank you for the review and the suggestion! I originally started with an exponential distribution but the varying growth/decay rates led me to shift to a linear curve, thats a good spot! I will be incooperating day-level analysis in the coming posts and I sure hope this turns out to be accurate!

      Like

  6. Nice way of using predictive analytics. Prediction of 4.91 lakhs cases by 5th July looks believable unlike some big universities and statistical organizations unrealistically predicting India will have 10 crore cases by July Mid. Keep going. One suggestion from my side. If you can explore your predictive analytical skills in predicting stock market price movements that would be a great thing to enhance your predictive analytical skills.

    Liked by 1 person

    1. Thank you for the review and suggestion! Most articles often exaggerate numbers so I tried to keep it as fair as possible. I will be looking forward to enter that sphere of predictive analysis in the mid-term future, stay tuned!

      Like

  7. Amazing Research.. Really liked the fact that you have been able to pin down the micro factors location and region wise in addition to the nation wide scenario. This is important as almost every city has its own local challenges, culture, population density and how they are handling the covid-19 situation. Also may be good idea to re-train the model in June given various Govt directives will be in place and will allow latest growth rates and impacts to be reflected in the models.

    Liked by 1 person

    1. Thank you! The model is currently a very basic version, I’ll be working on it to include more ML concepts for a better output or analysis for different kinds of situations.

      Like

  8. Dear Aaryan,
    Splendid job, I am not really sure if we had that much capabilities at our disposal when we were your age however if you are inclined towards finding the truth via stats and mathematic equations then believe me we have scientists writing/deriving equations to govern/replicate each and every action plausible within nature to contribute to robotics.
    Your article is really good and gives an analysis in detail at high level but it made me think can there be more unknown parameters to be considered into the formulation of factors of x and y and even inside of it if there are other variables which ought to be considered like population behavior, medical amenities, mental setup etc in Delhi and comparison to that of Mumbai, MH. Its just like differentiation of a circle in points parallel to a line will yield a line but if you integrate those points about a point you can achieve the circle back yet it comes with certain constants in the equations. Might I suggest that you pursue the same approach with Neural Network algorithms and treat two different cities like Delhi and Mumbai for instance to compare on similar grounds and identify KPIs which can be more specific to the life in those cities like e.g. rate of growth of cases[reported] across lockdowns, rate of recovery from reported cases, probability of existence of unknown cases[numbers] etc. I understand you have gone down to the case of South Delhi specifically however here is the thing … to build a constant with an equation, the relatability of the equation and variables to the ecosystem is also necessary however for now to derive a generic balance between locations within the cities in general might provide good insight directly. Just a few ideas, I thought you might want to pursue to let the system identify and co-relate the unknowns based on sample set available.
    This has been quiet insightful and I have been asking all the websites hosting COVID status with more KPIs and information but none believe in bringing the intelligence out of the existing numbers or perhaps they see something more horrifying in the future … 🙂
    Nevertheless pursue the course of truth via mathematics … its a great science … Kudos to you and everyone involved here.

    Like

  9. It’s a very informative read Aaryan, The predictable analysis seems to be very realistic and all the calculations are very well in place and hope this pandemic situation winds up soon. Also liked the overall theme with embedded features and plugins used on the website. Good Job !

    Liked by 1 person

  10. Mathematical Models and Technological content are very precise and impactful.
    Well articulated research document.

    You are on the right path Aaryan!!

    Liked by 1 person

  11. Nice job Aryan! the contents are very precise, you are on right path. It will be a good idea to re train the model with latest details in upcoming months with latest growth rate and government directives.

    Liked by 1 person

  12. Excellent work Aaryan, you have built a pretty good model to calculate the trend and predict future numbers, the model reflects near real world scenario, although real world is more complicated and hard to consider, for example some events(big congregation, government interventions etc) which will impact the prediction. Nice job Aaryan!

    Liked by 1 person

    1. Thank you! I tried my best to keep it as simple and accurate as possible, but will surely be adding new methods to it to make it more accurate in a wide variety of scenarios.

      Like

  13. Some exceptional work there, Aaryan. Lot of good feedback already from some of the best people. All the best for your next.

    Liked by 1 person

  14. Your rsearch paper is quite descriptive with many details.You can try enhancing and adding more attributes to your predictive model. This might help you in getting more precise predictions.
    Well done. Long way to go. Kudos 🙌

    Liked by 1 person

  15. Good job Aaryan. Its very well articulated. You can also think of adding the impact of migrant population and their future impact on a particular area when they start arriving from infected zones. Also this disease have multiple waves and the unique case of Singapore (succeeded in first wave and faltered in second) will be a good scenario to refer.

    Liked by 1 person

  16. Very Impressive. Really liked the in-depth analysis of the content. Keep it up!!!
    Looking forward for the upcoming Football topic.

    Liked by 1 person

    1. Thank you! Being a football fan myself, it is always on the table for topics to write on, once the season resumes more football related posts will be out.

      Like

  17. I would like to congratulate you Aaryan for writing this research paper and backed it with predictive analytics. Very well articulated and super informative. Keep up the good work !

    A suggestion from my side, if you think if it make sense then add a section at the end of the paper as ‘Conclusion : What India needs to do right now to overcome with COVID-19’. You can refer the below article published by AIIMS doctor where simple measures are provided at the end of article to overcome this pandemic.
    View at Medium.com

    Liked by 1 person

    1. Thank you for the review, help and the suggestion! I am currently working on a sequel to this paper, that will include these details, stay tuned!

      Like

  18. Well Done Aaryan !! Keep up the good work 🙂 You are doing a great job. The model is helpful in many ways for the country and for your career growth as well. Keep the momentum going.

    Liked by 1 person

  19. Very well Articulated Analysis. Exploring the facts and collecting data points helped to make it Empirical. Kudos ! Your effort is evident.

    Food for thought – Please gauge the accuracy of model against actual number. This will help to make any corrections, if needed. Wish you get max accuracy!

    Liked by 1 person

    1. Thank you for the review and suggestion! I included comparisons in figures 2,4 and 8 but I will be adding more details to the predictor for better results.

      Like

  20. Good Analysis and very well articulated facts. Just a suggestion, if you can include What next should be done in India to reduce the numbers? you can refer to the paper by AIIMS regarding the same.

    Liked by 1 person

  21. Well researched article Aryan! You have covered almost all the parameters. Statistics and inferences look believable…just a minor point – bullet your inferences, that will help increase focus and interest of readers. All in all – great start !

    Liked by 1 person

  22. Model is very innovative and well-articulated. Appreciate your motivation and enthusiasm to put this together on subject area which is so relevant in today’s global situation. Great start and very well done Aaryan! Keep them coming.

    Liked by 1 person

  23. Hi Aaryan, Many congratulations on coming up with such a well-articulated and data driven analysis!

    Its exciting to see a young mind like you so passionate to explore the power of science behind data and to be actually able to implement that into live scenarios and come up with answers to some of the most relevant questions in today’s time.

    Keep up the enthusiasm and good work !

    Liked by 1 person

    1. Thank you for the detailed review! I will be coming up with more posts related with the use of data science in the near future, stay tuned!

      Like

  24. Good work Aaryan.
    I liked your inference on the red zone disparity particularly.
    Coming to the growth rate predictor, I ran the REPL file and would like to know the numbers you used to get the growth rates post lockdown. Using a linear function to get the recovery rate is nice, but don’t you think it is better to use exponential function to get the growth rates post lockdown?
    Waiting for your paper on football.
    You’ll never walk alone 😉
    Cheers

    Liked by 1 person

    1. Thank you! I proceeded to use the linear growth function rather than the exponential one as the rates were very variable and I was able to get more accurate results with the linear model. Football content coming very soon!

      Like

  25. It is really a nice analysis on today’s most important challenge of society .I believe Analytics & AI has a role to play in fighting this pandemic and hope to soon put our system to work helping whoever working on the front lines.

    Liked by 1 person

  26. The comprehensive Analysis, Projections, Predictions and use of Advanced Technologies shows you are Next Generation Technology Leader.
    There will be endless corrections for even a perfect model. There will be always an upgrade and versions. So take the suggestions but don’t bother now you are the best and tomorrow best.1.
    All the best!

    Liked by 1 person

    1. Thank you! You’re absolutely right, the suggestions will be used in new posts or for minor corrections that need to be made in the original post.

      Like

  27. Good work Aaryan. Your website is well put together, it has a very clean interface and the information in on point. Keep up the good work.

    Liked by 1 person

Leave a Reply to Vikrant Mahajan Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: