We at MachineHack feel it is crucial to predict when the outbreak may slow down, flatten or further worsen across different countries to ascertain the true nature of economic or human life cost as a result of COVID-19.
We are in the midst of a worldwide outbreak of respiratory illness induced by a novel coronavirus, initially identified in China and which has now been detected in more than 110 geographies internationally.
On 7th March 2020, the World Health Organisation (WHO) also proclaimed that the cumulative number of confirmed cases for COVID-19 had surpassed 100,000. The organisation urged all countries to continue their efforts to curb the disease, which has expanded to become a global pandemic.
The virus outbreak has become one of the biggest threats to the global economy and financial markets. Major institutions and banks have cut down on their growth forecasts and the stock market has seen a drastic plunge worldwide.
In this context, we feel it is crucial to predict when the outbreak may slow down, flatten or further worsen across different countries to ascertain the true nature of economic or human life cost as a result of COVID-19.
The Objective Of The Hackathon
In the coming weeks and months, we at MachineHack (an Analytics India Magazine initiative) along with our community members will ominously examine how the coronavirus could affect different nations.
Thereby, we invite MachineHackers to predict potential COVID-19 cases across all the globe on an everyday basis. The objective of the hackathon is to gauge COVID-19 on three metrics- confirmed cases, recovered cases and death events for the next day using historical data as on a given date.
As sad as it is to analyse the data around COVID-19 events, it is critical to keep a tab on the disease metrics to track the outbreak. The hackathon will be based on the data published by various agencies and the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE), which can be found below-
Find the JHU CSSE Github Repo here.
A Note For Hackathon Participants
The univariate time series knowledge is rendered based on all individual countries affected by COVID 19 from 22nd January 2020 onwards to the present date.
Here is an example for your reference. The provided .csv file comprises the count of confirmed COVID cases across countries till 10th March 2020 for the three target variables (confirmed cases, recovered cases and death cases).
The dataset would be updated daily at 00:00 UTC standard time with the prevailing forecast of the distinct target variables. It is to be noted that the published data is dynamic, and hence it will be renewed each day in a new column every day. The data in the rows will also fluctuate based on the reported changes for COVID-19 outbreak in various world geographies.
The submission file from participants must contain the projected count of incidents for the next day, i.e. 11th March 2020. as per the sample_submission.xlsx format.
Instructions For Participants
Participants must submit a solution every day to avoid penalty in the final leaderboard standings at the end of the hackathon
The leaderboard for this hackathon is not dynamic and will only update once between 7 am and 8 am (Indian Standard Time) every day
All invalid submissions will receive an error code of 12345 on the leaderboard, and will not be considered for determining the final leaderboard standings at the end of the hackathon
Participants can submit only one solution per day
All current day submissions from users will be evaluated against the next days actual report as per Johns Hopkins University.
Country/Region: Name of Country/Region.
Date Stamp: The sequence of historical counts with since 22nd Jan 2020.
The 3 .csv files contain historic counts per country for (Confirmed, Recovered and Death).
1. covid_confirmed_daily_updates.csv – Contains the count of confirmed COVID cases.
2. covid_deaths_daily_updates.csv – Contains the count of COVID patient deaths.
3. covid_recovered_daily_updates.csv – Contains the count of recovered COVID patients.
4. sample_submission.xlsx – submission format for the model evaluation.
The datasets are dynamic and will be automatically updated at 00:00 UTC standard time every day with the latest count of the respective target variables.
No Reviews found for this hackathon.