:pill: Ersilia -- GDI Data for Good Hackathon :rocket:

The repository for Hackathon related information!

:pushpin: Table of Contents

• :sunny: Overview
• :calendar: Key Dates
• :link: Key Links
• :dart: Problem Statement
• :bar_chart: Dataset
• :sparkles: Expectations
• :trophy: Judging Criteria
• :computer: GDI Hackathon: Format
• :busts_in_silhouette: Teams

:sunny: Overview

This repository contains the data and code for the Ersilia Hackathon. The goal of the hackathon is to use data to help Ersilia, a non-profit organization, better understand the needs of their clients and improve their services.

:calendar: Key Dates

• Kick-off Call: Monday, 20th May 2024
• Midpoint Check-in: Monday, 17th June 2024
• Presentation: Tuesday, 9th July 2024

:link: Key Links

• Ersilia Hackathon Slack Channel
• Ersilia Website
• Hackathon Kick-off Call Slides
• Ersilia GitHub
• GDI <> Ersilia Project Blog

:dart: Problem Statement

Background

Ersilia is a nonprofit organisation focused on lowering the barrier to drug development by providing computational tools for open-source drug discovery. The organisation specialises in building and deploying hundreds of predictive models for target profiling, drug repurposing, and safety evaluations, focusing strongly on neglected tropical diseases that affect the global south.

The challenge is three-fold:

1. GDI cohort volunteers will be involved in defining and developing ideas for “Impact Metrics” that Ersilia can use
2. Automating the collection of said "impact metrics".
3. Presenting the impact metrics in an innovative and novel way via a data product of your choice :muscle:.

Current approach deployed by Ersilia for impact reporting:

• Content database (spreadsheets) compiling publications, code repositories, funding raised, etc. No automated dashboard is available.
• For global health statistics such as disease burden or investment on R&D in the global south, Ersilia manually searches the scientific literature.

:bar_chart: Dataset

The datasets involved will include:

Internal spreadsheets detailing:

• Code repositories, contributions, and activity. This includes the models developed and other tools of interest to drug discovery.
• Scientific publications and blogposts detailing outreach activities.
• Community members, with a focus on mentorship in the global south.
• Events and presentations, including capacity-building workshops.

Externally available datasets are recommended to broaden the scope of the impact reporting, such as:

• WHO Global Health Data.
• Our World In Data.
• Open Alex.

Workflows used by Ersilia to collect internal stats, for inspiration

• Ersilia Stats repository.

:sparkles: Expectations

• Defining and calculating crucial impact metrics for Ersilia.
• Developing automation scripts to collect and process these metrics efficiently.

Suggested recommendation topics and ideas:

• A memo outlining the definition of each metric (with a “glossary”).
• A live dashboard that Ersilia can present said metrics to prospective funders, donors, and the general public.
• Focus on the UN's Sustainable Development Goals (SDG).
• Automation workflows, ideally in GitHub Actions, to collect information from external resources.

:hourglass_flowing_sand: Mid-point check in

• For the mid-point check in, be ready to share your answers (<5 mins) with the group
• No slides are required :smile:

1. What have we achieved so far?
2. What are the next steps?
3. What questions do you have for Miquel and Dhanshree? -- We'll note down everyone's questions and have Miquel and Dhanshree answer them at the end :smile:
4. What's been your biggest success so far?
5. And biggest a challenge?

:trophy: Judging Criteria

Practical/Actionable Insights (40%)

• Covered a range of relevant considerations with rigour, thoughtfulness.
• The logic was clear and correct, and the approach was structured.
• Insights are actionable.

Relevance to the given problem statement/understanding of client issues (40%)

• Clearly defined the client's problem and issues.
• Insights are relevant and impactful to the client's context.

Final Presentation (10%)

• The presentation was clear, well-structured, and engaging.

Creativity/Originality (10%)

• Applied a new, different, and creative approach, pushing beyond the obvious.

:computer: GDI Hackathon: Format

The hackathon is an open-ended format - unlike a traditional GDI project with clearly defined scope and deliverables, in the Hackathon, we want you to flex your data-science muscles and show us what impact you’re able to have!

We find that everyone will approach the challenge in different ways. This is valuable for our charity partner(s) because we expect to effectively come up with four different solutions and four different perspectives.

In terms of tools, anything goes! We have found in hackathons that people take this in different directions because they have different skill sets. Some of you will have a business intelligence background and skills in dashboarding segmentation. Others will have come from more technical backgrounds and have skills in data engineering, modelling, cloud, etc. Use what makes the most sense to you and leverage your team's strengths!

The main source of news/announcements and other updates throughout the Hackathon will be via the GDI Slack Channel titled #hackathon_ersilia. If you have any questions feel free to shoot them in this Slack channel or directly reach out to Berta, Shakeel or Vivek.

This channel will be an open forum to bounce ideas around and ask questions (please don’t hold back!).

:busts_in_silhouette: Teams