Data Scientist Consultant – Climate Anomalies

Invitation to bid.

BACKGROUND

The Internal Displacement Monitoring Centre (IDMC) is the world’s definitive source of data and analysis on internal displacement. Since our establishment in 1998 as part of the Norwegian Refugee Council (NRC), we have offered a rigorous, independent and trusted service to the international community. Our work informs policy and operational decisions that improve the lives of the millions of people living in internal displacement, or at risk of becoming displaced in the future.

With a team of almost 40 people in Geneva, we monitor more than 252 countries in the world for which we provide verified, consolidated and multi-sourced estimates of the number of people internally displaced or at risk of becoming displaced by conflict, violence, disasters. Our data and evidence are published on our website and via our Global Internal Displacement Database (http://www.internal-displacement.org/database/displacement-data). Our flagship report, the Global Report on Internal Displacement (https://www.internal-displacement.org/global-report/grid2024/), *published every year in May, is the world reference on IDP statistics, featuring trends and thematic analyses, country and regional spotlights, including maps.

IDMC is looking for data scientist to generate new insights on climate anomalies attribution. The primary objective of the work is to enhance our data on displacement associated with weather-related hazards. This involves adding metadata to events that trigger displacement, particularly those identified as climate anomalies sourced from NOAA. This enhancement aims to provide a more comprehensive understanding of climate change impact on disaster displacement. For this task, the consultant will work with the IDMC NLP data extraction pipeline which is a python package designed to extract structured data from documents.

CONSULTANCY OBJECTIVE

Under the supervision of the Information Management Adviser in the Data Management and Risk Analysis Hub, the consultant will develop the “NOAA crawler” class to systematically identify and download relevant climate anomaly documents and datasets from NOAA global monthly reports.

The consultant will identify and define the most relevant fields related to climate anomalies for integration into the IDMC internal database. This includes conducting meetings with stakeholders, documenting the rationale behind field selection, and preparing a schema for database integration.

Based on the above steps, the consultant will develop a “NOAA extractor” class utilizing the IDMC NLP pipeline for extracting key data points (dates, locations, descriptions) from NOAA reports.

Lastly, the consultant will be responsible for ensuring the matching of the extracted climate anomaly events with existing disaster events in the HELIX database. This involves developing or utilizing algorithms for event matching based on criteria such as date, location, and event type. Rule-based, statistical, and LLM matching methods should be explored.

The consultant will also compile the project’s findings into a comprehensive paper or blog post for public dissemination. This involves analyzing the data and insights gathered, writing the paper or post, and publishing it on relevant platforms.

CONSULTANCY TIME and IMPLEMENTATION TIMEFRAME

This is a consultancy of approximately twenty five (25) days over the course of four (4) month from approx. 15th June to 15th October 2024.

This is a home-based consultancy, with regular interactions through teleconferences or face-to-face meetings.

Key Deliverables include:

• Development of a “NOAA crawler” class to systematically identify and download relevant climate anomaly documents and datasets from NOAA global monthly reports.
• Identify and define the most relevant fields related to climate anomalies for integration into the HELIX database.
• Implement the “NOAA extractor” class utilising the IDMC NLP pipeline for extracting key data points (dates, locations, descriptions) from the NOAA documents. Validate and report challenges across a wide set of documents.
• Match the extracted climate anomaly events with existing disaster events in the HELIX database. This involves developing or utilising algorithms for event matching based on criteria such as date, location, and event type. Ruled-based, statistical, and LLM matching should be explored.
• Compile the project’s findings into a comprehensive paper or blog post for public dissemination. This involves analysing the data and insights gathered, writing the paper or post, and publishing it on relevant platforms.

MANAGEMENT

The consultant will be line-managed by the Information Management Adviser within the Data Management and Risk Analysis Hub.

BUDGET and REMUNERATION

Applicants are expected to provide a daily fee inclusive of all taxes and charges for this work.

SKILLS and QUALIFICATIONS

• Minimum graduate degree in a relevant field such as data science or computer science.
• 4+ years working and/or studying in the field of data science and/or statistics.
• Exceptional skills in Python, with a strong grasp of advanced object-oriented programming concepts.
• Proven track record developing data and analysis processes and systems linked to effective decision support functions.
• Proficiency in using GitHub for version control and collaborative projects.
• Natural Language Processing (NLP) Expertise: Demonstrated experience in NLP projects.
• Experience with Large Language Models (LLMs): Hands-on experience with LLMs.
• Climate Science Knowledge: A good understanding of climate attribution science and experience working with disaster impact data.
• Fluency in English; proficiency in additional languages such as French or Spanish is advantageous.
• Excellent communication and writing skills in English.
• Excellent problem solving and analytical skills.
• Ability to deliver according to agreed deadlines.
• The consultant must be a registered company or a sole trader as required in country of residence.