top of page
Post: Blog2_Post
Writer's pictureStephen Turban

3 BitGrit Competitions You Need to Check Out as a High Schooler

If you're a high school student with an interest in AI or data science, participating in AI-based competitions can help you expand your knowledge, improve your skills, and challenge yourself in ways that traditional coursework may not. While traditional education provides a strong foundation, competitions offer a chance to apply theoretical knowledge to real-world situations.

 

Taking part in data science competitions can improve your coding skills, enhance your analytical abilities, and allow you to tackle complex challenges similar to those faced in the industry. These experiences can also help you develop important skills such as critical thinking, problem-solving, and effective communication, which are valued by college admissions officers. 


Additionally, your competition achievements will strengthen your portfolio, which can be included in your college applications, showcasing your passion for the subject and your commitment to learning, potentially setting you apart from other applicants.


In this blog, we'll highlight three BitGrit competitions worth exploring to deepen your interest in AI.


Why BitGrit?


Platforms like BitGrit provide a great setting to gain hands-on experience in the field of artificial intelligence. With a range of competitions designed for various skill levels, BitGrit will allow you to engage with challenges that align with your interests, whether those are in environmental science, healthcare, or space exploration. These competitions will connect you with a global community of data science enthusiasts, fostering collaboration and growth. Whether you’re a beginner or an expert, BitGrit’s platform offers the tools and opportunities to elevate your skills and make a tangible impact in the field.


To help you get started, we’ve compiled a list of currently available BitGrit competitions specifically for high school students.


1. Bird Species Classification Challenge (Learning Competition)


This is a learning-based competition with no monetary prize, where you will focus on gaining knowledge and building your skills.


As we learn more about the world, taxonomists work to identify and describe the species that inhabit it. Each year, thousands of new species and subspecies are discovered and cataloged, sometimes requiring reclassification based on unique characteristics.


Once identified, tracking and estimating species populations is crucial—a task made more efficient by machine learning. So, the main aim of this challenge is to develop an ML model to predict bird species using their attributes and geographical locations.


Problem Statement:

Scientists have reclassified a known bird species into three distinct species, all endemic to a specific region. A non-profit conservation society needs to track and log these species based on field observations. Using genetic traits and location data, your goal will be to predict the observed bird species (A, B, or C) in this beginner-level practice competition.


Datasets:

This competition includes two datasets: training datasets and test datasets. Both contain bird data from locations 1 to 3. The training_set can be joined with the training_target using the 'id' column. The objective is to develop an algorithm that predicts the "Species" in the training_target.csv.


Rules for submission:

  • The submission file must match the format of the example file, solution_format.csv.

  • Submissions are evaluated on accuracy (Number of correct predictions / Total number of predictions).

  • You are allowed to submit a solution file up to three times per day.

  • You can submit up to 3 entries per day. Additional submissions must wait until the next day.

  • External datasets are not allowed.

  • Do not upload the competition dataset to other websites.

  • All submissions must be made individually, as team submissions are not permitted.

  • The leaderboard is rolling; submissions older than 90 days won't count.

For further inquiries, you can contact them at info@bitgrit.net.

Start Date: May 18, 2022

End Date: This is an ongoing competition. 

Prize: This is a learning competition. Apart from gaining knowledge, there are no prizes for this competition.


2. Weather Forecast Challenge (Learning Competition)


This is a beginner-level competition focused on predicting the next day's weather using prior-day data. It has no monetary prize.


Weather forecasting has always been crucial for survival, impacting everything from basic needs to major industries like construction and transportation. However, predicting the weather is complex due to the massive amount of data involved.


In agriculture, precise weather forecasts are vital. Farmers rely on accurate predictions to make decisions that affect crop health and yield.


Problem Statement:

A major agricultural company needs an accurate weather prediction algorithm to optimize growth, save resources, and improve production. Using historical weather data from neighboring regions (A-E), you will have to predict the next day's weather for the target area (N: No rain, L: Light rain, H: Heavy rain). 


Datasets:

This competition includes two datasets: train datasets and test datasets. Both contain weather data for regions A through E. Note that regions A through E and the target region for weather prediction are all neighboring areas, but the specific location of each region is not provided. The datasets can be joined with the solution file using the 'date' column. Your task will be to develop an algorithm to predict the "label" in the solution_format.csv. Please be aware that all values in solution_format.csv are placeholder values.


Rules for submission:

  • The submission file must match the format of the example file, solution_format.csv. The file should contain two columns: one for 'id' and one for 'value'. Both columns must be formatted as strings, meaning each entry should include double quotation marks ("). If the values are submitted as numeric without quotation marks, the score will be 0.

  • Submissions are evaluated based on accuracy, which is calculated as the number of correct predictions divided by the total number of predictions.

  • You may submit a solution file up to three times per day.

  • Up to 3 submissions per day; additional submissions must wait until the next day.

  • External datasets are prohibited.

  • Do not upload the competition dataset to other websites.

  • Submissions must be individual; no teams are allowed.

  • The leaderboard is rolling; submissions older than 90 days won't count.

For further inquiries, you can contact them at info@bitgrit.net.

Start Date: March 22, 2022

End Date: This is an ongoing competition. 

Prize: This is a learning competition. Aside from gaining knowledge and building skills, there are no prizes for this competition.


3. The NASA Breath Diagnostics Challenge (Prize Money Competition)


NASA's Science Mission Directorate offers students an opportunity to enhance the accuracy of the NASA E-Nose, a potential clinical tool for diagnosing diseases by analyzing human breath. Your task will be to develop a classification model that can distinguish between COVID-positive and COVID-negative breath samples using data from a recent clinical study. 


The challenge aims to develop a diagnostic model using NASA E-Nose data from the exhaled breath of 63 COVID-19 study volunteers. You will apply advanced AI and data preparation techniques to overcome the study's small sample size. Successful solutions could enhance the NASA E-Nose's capabilities for clinical applications in human space exploration.


This competition is designed for participants with advanced skills in data analysis and artificial intelligence.


Datasets:

The data consists of 63 text files, each representing one of 63 patients, numbered from 1 to 63. Each file includes the Patient ID, COVID-19 diagnosis result (POSITIVE or NEGATIVE), and numeric measurements from 64 sensors (D1 to D64) installed in the E-Nose device, which detect molecular signals in patients' breath.


Rules for participating in this competition are as follows: 

  • You are allowed to submit up to two entries per day. Any additional submissions must be made the following day. Attempts to exceed this limit will result in disqualification.

  • External datasets are prohibited. However, you may use derivative datasets like calculated features and synthetic training data.

  • Do not upload the competition dataset to other websites; doing so will lead to disqualification.

  • You must manually select your final submission (up to 2) before the competition ends. Otherwise, the one with the highest public score will be automatically selected.

  • In case of a tie on the private leaderboard, the participant who submitted first will advance to the next stage.

  • After the competition, top participants are required to submit the necessary files and documentation by September 16, 2024, for the final review. Failure to meet this deadline may lead to disqualification.

  • Your submitted solution must reproduce the same model and inferencing output as shown on the leaderboard. Ensure the random state is set appropriately for reproducibility.

  • Exploiting non-statistical patterns, anomalies, or data artifacts will result in immediate disqualification.

  • Models must perform inference efficiently on standard consumer-grade hardware within a reasonable time frame.

  • Scores will be validated internally. If the Internal Score deviates by more than 10% from the Overall Score, the Internal Score will be used as the Final Score.

  • Prizes will be awarded based on the Final Score and are subject to eligibility verification. The competition winner must transfer all rights to NASA, including copyrights, patents, and know-how.

  • All decisions by bitgrit are final. Prizes may be subject to tax reporting and withholding requirements.

For further inquiries, you can contact them at info@bitgrit.net.

Start Date: July 5, 2024

End Date: September 6, 2024

Prize: The total prize pool for this competition is $55,000. For a detailed breakdown of the prize money allocation, you can check here.

The winners of “The NASA Breath Diagnostics Challenge” will be declared on November 8, 2024.


Another competition, the NASA Airport Throughput Prediction Challenge, began on September 13, 2024. This competition is open to students who are 18 years old or older and are affiliated with a U.S. university.



Stephen is one of the founders of Lumiere and a Harvard College graduate. He founded Lumiere as a PhD student at Harvard Business School. Lumiere is a selective research program where students work 1-1 with a research mentor to develop an independent research paper.


bottom of page