Menu

TMDU-MIT-NUS-ANZICS-JSICM Critical Data Workshops and Datathon 2023

Overview of the critical data workshops

Critical data workshop 0: Preparatory materials

click here for further information

Critical data workshop 0: Preparatory materials
The preparatory workshop provides crucial background information about the inaccuracies of pulse oximeters for populations with darker skin tones. It explains how these inaccuracies can lead to missed cases of hypoxemia, resulting in less treatment and higher mortality rates. The workshop also provides an overview of the Datathon's schedule and recommends literature for the participants to review. Additionally, the participants are prepared for the data analysis.

Critical data workshop 1: EDA & study design

click here for further information

Critical data workshop 1: EDA & study design
In this workshop, the teams are expected to define the inclusion criteria to build a working cohort and explore the ground truths present in the dataset. The deliverables include a flow chart detailing the inclusion and exclusion criteria and the definition of the machine learning task to be addressed. The workshop also discusses potential pitfalls such as sampling and representation bias.

Critical data workshop 2: Clinical variables selection & feature engineering

click here for further information

Critical data workshop 2: Clinical variables selection & feature engineering
This workshop emphasizes the importance of data preprocessing in the data science workflow. Participants are taught data cleaning, normalization, transformation, and reduction techniques, and they apply these techniques to real-world datasets. The workshop also highlights how to identify and mitigate biases that can be introduced during data preprocessing.

Critical data workshop 3: Let's get our model!

click here for further information

Critical data workshop 3: Let's get our model!
This workshop involves splitting the data into training and test sets, defining performance metrics for model evaluation, and developing a machine learning model. The models could predict arterial oxygen saturation values, the gap between arterial and peripheral oxygen saturation, or detect instances of hidden hypoxemia. The workshop also covers grid-search parameter tuning and feature importance assessment.

Critical data workshop 4: Try to tackle the biases and re-model

click here for further information

Critical data workshop 4: Try to tackle the biases and re-model
In the final workshop, participants are encouraged to develop their own ideas and solutions to tackle biases and remodel the datasets based on the knowledge and skills acquired from the previous workshops. It promotes creativity and self-guided exploration and encourages participants to address various problems, such as demographic biases, missing patients from the database, developing a good fairness metric, and validating and expanding the proposed models.

What is Datathon

This event caters to physicians keen on exploring research from a data science perspective and data scientists seeking to leverage their skills to resolve clinical problems. No prior experience in both data science and clinical knowledge is required ? just an openness to learn .
Participants will have the chance to network and learn from leading professionals in data science and medicine. The event presents an excellent platform to learn, interact, and contribute to significant advancements in medical research.

Benefits for participants include:

- Learning opportunity: Gain practical experience in data science and understand its application in real-world medical scenarios.

- Expert guidance: Learn from leading academics and industry professionals from globally recognized institutions.

- Interdisciplinary collaboration: Collaborate with physicians, clinical researchers, and data scientists to solve actual clinical problems.

- Networking opportunities: Connect with peers, potential collaborators, and industry leaders.

- Impactful contribution: Apply new skills to real-world problems, potentially contributing to significant advancements in the medical field.

- Career advancement: Boost your skills and open new avenues in your professional or research career.

- Awards and recognition: Top three projects will be recognized, opening the door to potential future opportunities.

- Publication opportunity: Teams are encouraged to publish papers based on their findings.

By participating in this Datathon, attendees will be at the intersection of data science and medical research, potentially driving innovative solutions to real-world problems. This event represents more than just a learning experience; it's an opportunity to make a substantial difference in healthcare.

-学習の機会:データサイエンスの実践的な経験を積み、実際の医療シナリオへの応用を理解する。

-専門家の指導:世界的に著名な研究機関の一流の学者や業界の専門家から学ぶ。

-学際的コラボレーション:医師、臨床研究者、データサイエンティストと協力して実際の臨床問題を解決する。

-ネットワーキングの機会:同業者、共同研究者候補、業界リーダーと交流をもつ。

-インパクトのある貢献 新しいスキルを現実の問題に応用し、医療分野の大きな進歩に貢献する可能性がある。

-キャリアアップ:スキルを高め、専門家や研究者としてのキャリアに新たな道を開く。

-賞と表彰:上位3つのプロジェクトが表彰され、将来の可能性への扉が開かれる。

-出版の機会:各チームは、研究成果を論文として発表することが奨励されます。

このDatathonに参加することで、参加者はデータサイエンスと医学研究の交差点に立ち、現実世界の問題に対する革新的な解決策を推進できる可能性があります。このイベントは単なる学習体験ではなく、実践的なインパクトを与える機会とすることも推奨されます。