Data Science for Effective Good

SEADS is a Data Science consultancy focused on supporting highly-effective charities. Our services cover a broad range from executing concrete projects to using data for informing strategical decisions or optimizing data pipelines in your organization. We can help you with a broad range of projects, including:

  • Exploratory Data Analysis and Data Visualization
  • Survey analysis
  • Setting up dashboards and data pipelines
  • Data scrapping
  • Network analysis
  • Natural Language Processing
  • Computer Vision
  • Statistical modeling
  • … most of other data-related tasks

Even if you don’t have a concrete project in mind, but think that you could benifit from better interaction with your data, please reach out to us!

Our Portfolio

Water Data Analysis for the Fish Welfare Initiative

Exploratory data analysis for the Fish Welfare Initiative to inform fish farming policies in India. Investigation of how water quality (e.g. oxygen concentration) varied across fish farms, depending on the installed facilities, and influenced fish disease and population.

Designing a donation tool with ClearerThinking and GWWC

ClearerThinking developed a tool that helps people to understand their values and decide on the best way to maximize their positive impact. For Giving What We Can, our team analyzed user responses from this tool to obtain insights that could inform the development of a similar one for deciding on target charities for donations.

Analyzing view metrics on the EA Forum

Analysis of how the popularity of the EA forum changed over time and how views are distributed across posts and topics.

See the post for more details.

EAG survey data analysis

On behalf of CEA, SEADS analyzed survey data for EA Global events (as well as EAGx and EAG Virtual events) ranging from 2016 to 2020. We estimated value based on how likely articipants were to recommend future conferences. We concluded from our analysis that community-related variables (such as feeling welcome or making connections with new people) are more closely linked to this likelihood than content-related variables (such as enjoyment of speaker sessions). Our demographic analysis of community-related variables yielded no significant differences based on participants’ age, ethnicity, or gender.

See our post for more details.

TalkItOver chatbot evaluation

TalkItOver is a mental health chatbot inspired by the Samaritans’ methodology.

For this project we:

  • evaluated its expected impact
  • analyzed conversation data to determine the situations where the bot helps the clients and where it doesn’t
  • proposed potential improvements in the bot algorithm based on our findings
  • provided advice for the strategic development of the project

GWWC Influencer Network Analysis

Scraping of GWWC Twitter network and network analysis on most influential nodes. The aim is to improve the spreading of EA ideas.