Scribd
Upload
14 views6 pages

Hackathon Notes

The document outlines the tools and structure for a hackathon, including the development environment, core libraries, dataset format, and communication methods. It details judging rubrics for evaluating problem formulation, technical approach, results, and presentation, each with specific scoring criteria. The emphasis is on understanding the chosen problem, executing a sophisticated technical approach, deriving actionable insights, and effectively communicating results.

Uploaded by

Trusfrated Army
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
14 views6 pages

Hackathon Notes

The document outlines the tools and structure for a hackathon, including the development environment, core libraries, dataset format, and communication methods. It details judging rubrics for evaluating problem formulation, technical approach, results, and presentation, each with specific scoring criteria. The emphasis is on understanding the chosen problem, executing a sophisticated technical approach, deriving actionable insights, and effectively communicating results.

Uploaded by

Trusfrated Army
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Tools

●​ Development Environment: Google Colab (online Jupyter Notebook, etc.)


●​ Core Libraries:
○​ Pandas: For data manipulation and analysis.
○​ NumPy: For numerical operations.
○​ Scikit-learn: For implementing a wide range of machine learning models.
○​ Matplotlib & Seaborn: For data visualization to understand the dataset and
model results.
○​ PyTorch/TF
●​ Version Control & Collaboration: GitHub → Main repository includes the dataset,
starter code notebooks, and list of reference links. Each team can then fork this
repository.
●​ Communication: Dedicated Discord server or Slack workspace? Or maybe just email?

Dataset Format
●​ Format: CSV
●​ Dictionary: A README.md file that explains:
○​ What each column in the dataset represents.
○​ The data type of each column (e.g., integer, string, float).
○​ Any known missing values and how they are represented (e.g., NaN, null).
○​ The target variable that students need to predict.
●​ Starter Notebook: Provide a basic Google Colab:
○​ Load the dataset using Pandas.
○​ Build a very simple baseline model using Scikit-learn.
○​ Generate a submission.

Questions
●​ Provide multiple questions that could be answered and let students decided
○​ Choosing the important/impactful problem to solve is often more important than
simply having a good model.

Hackathon Judging Rubrics


Score = (0.30×Problem​)+(0.30×Technical​)+(0.25×Results​)+(0.15×Presentation​)

Rubric 1: Problem Formulation & Justification (Weight: 30%)


Assesses how well the team understood/selected a question from the provided list, and
how convincingly they articulated its potential value (why does this question matter).
Score Description

1 (Needs The chosen question is not clearly stated or misinterpreted. Little to no


Significant justification for why this specific question was chosen or its real-world
Improvement) relevance. Shows minimal understanding of how their proposed work
will answer the question or derive value.

2 (Fair) The chosen question is identified but their understanding of it is


somewhat superficial. Justification for its selection or potential value is
weak. Demonstrates a basic, but not comprehensive, understanding of
how their solution will address the question.

3 (Good) The chosen question is clearly stated and understood. Reasonable


justification for its selection and its relevance. Shows a good
understanding of how their project will contribute to answering the
question and what value might be derived.

4 (Very Good) The chosen question is clearly and precisely articulated, demonstrating
a thoughtful selection from the given list. Strong and convincing
justification for why this particular question is important and relevant to
the dataset. Shows a deep understanding of how their approach will
directly answer the question and deliver significant potential value.

5 (Outstanding) The chosen question is exceptionally well-understood and presented,


highlighting its profound relevance and potential impact from the
provided list. The justification for its selection is compelling, insightful,
and clearly articulates how their proposed solution will not only answer
the question but also generate substantial, innovative value. The team
demonstrates strategic thinking in their choice and framing.

Rubric 2: Technical Approach & Execution (Weight: 30%)


This rubric evaluates the appropriateness, sophistication, and execution of the machine learning
models, data preprocessing, and overall experimental design.
Score Description

1 (Needs The technical approach is ill-suited to the problem or data. Significant


Significant errors in data preprocessing or feature engineering. Model selection is
Improvement) inappropriate or basic. Code is disorganized, contains major bugs, or is
not reproducible. Shows a lack of fundamental understanding of ML
principles.

2 (Fair) The technical approach is somewhat appropriate but lacks


sophistication or depth. Some issues with data preprocessing or feature
engineering. Model selection is basic or could be improved. Code is
partially organized but may have minor bugs or limited reproducibility.
Demonstrates a limited understanding of ML concepts.

3 (Good) The technical approach is appropriate for the problem, with generally
sound data preprocessing and feature engineering. Model selection is
reasonable. Code is mostly organized and reproducible, though minor
improvements could be made. Demonstrates a good understanding of
core ML principles and their application.

4 (Very Good) The technical approach is well-chosen and sophisticated, demonstrating


strong understanding of relevant ML techniques. Excellent data
preprocessing, thoughtful feature engineering, and robust experimental
design. Model selection is appropriate and shows consideration of
alternatives. Code is clean, well-documented, and highly reproducible.
Demonstrates strong technical skills and effective problem-solving.

5 (Outstanding) The technical approach is innovative, highly sophisticated, and perfectly


tailored to the problem and data. Exceptional data preprocessing,
creative and impactful feature engineering, and rigorous experimental
design. Demonstrates advanced knowledge of ML techniques,
potentially including novel applications or methodologies. Code is
exemplary: efficient, elegant, meticulously documented, and perfectly
reproducible. Shows mastery of technical execution.
Rubric 3: Results & Insights (Weight: 25%)
Assesses the quality and interpretability of the model's performance, the actionability of derived
insights, and the clarity of findings.

Score Description

1 (Needs Model performance is poor or not presented. Few to no meaningful


Significant insights derived from the analysis. Results are confusing, contradictory,
Improvement) or lack clear interpretation. There is no clear connection between the
results and the initial problem.

2 (Fair) Model performance is presented but may be unconvincing or


uninterpreted. Some basic insights are presented, but they lack depth or
actionable implications. Results are somewhat clear but could be better
explained or visualized. Limited connection between findings and the
problem's value.

3 (Good) Model performance is adequately presented and interpreted. Relevant


insights are derived, offering some value. Results are generally clear
and well-organized, with reasonable visualizations. A clear connection
exists between the findings and the proposed problem, though the
impact might not be fully explored.

4 (Very Good) Model performance is strong, clearly presented, and thoroughly


interpreted with relevant metrics. Insights are clear, actionable, and
contribute significantly to understanding the problem. Results are
well-visualized and compelling, providing a strong basis for
decision-making. The connection between the findings and the
problem's potential value is clearly articulated and convincing.
5 (Outstanding) Model performance is exceptional, rigorously validated, and presented
with profound interpretability. Insights are highly original, actionable, and
unlock significant new value from the data, potentially leading to novel
solutions or strategies. Results are presented with extreme clarity,
impactful visualizations, and provide definitive answers or strong
recommendations. The findings demonstrate a clear, significant, and
innovative solution to the chosen problem.

Rubric 4: Presentation & Communication (Weight: 15%)


Evaluates the team's ability to clearly present their work, explain complex concepts, and engage
with the audience.

Score Description

1 (Needs Presentation is disorganized, unclear, or incomplete. Key elements of


Significant the project are missing or poorly explained. Communication is difficult to
Improvement) follow, with poor explanation of technical concepts. The team struggles
to answer questions or convey understanding.

2 (Fair) Presentation has some structure but lacks flow or clarity. Some
important aspects of the project are not fully communicated.
Explanation of technical concepts is basic or somewhat confusing. The
team answers some questions but struggles with deeper inquiries.

3 (Good) Presentation is structured and mostly clear. All key elements of the
project are presented, though some details might be rushed. Technical
concepts are explained adequately for the audience. The team answers
most questions reasonably well.

4 (Very Good) Presentation is well-structured, clear, and engaging. All key elements of
the project are communicated effectively and concisely. Technical
concepts are explained clearly and accessibly to both technical and
non-technical audiences. The team handles questions confidently and
articulately, demonstrating strong understanding.

5 (Outstanding) Presentation is exceptionally well-crafted, highly engaging, and


perfectly paced. The narrative is compelling, and the project's value is
communicated with maximum impact. Complex technical concepts are
distilled into crystal-clear, intuitive explanations. The team demonstrates
mastery in answering all questions, engaging in insightful discussions,
and inspiring confidence in their solution.

You might also like