Skip to main content
SearchLoginLogin or Signup

Shared Philanthropic Data Schema: A Case Study with Lever for Change and Solve

In this publication, we explore our process for finding a shared schema of applicant data amongst our partner organizations. This includes the methodology we used for finding the schema, the queries for aligning the data, and the descriptions of shared columns (and near misses)

Published onAug 30, 2023
Shared Philanthropic Data Schema: A Case Study with Lever for Change and Solve
·

Methodology for finding a shared schema (administrative)

In order to find a shared schema between Lever for Change (LFC) and MIT Solve, each group was first instructed to follow our onboarding experience:

  • Fill out a data sharing agreement template

  • Philanthrobotics fills out a 2-team data sharing agreement meeting template, deciding on definitions and what data will be shared

  • Data is shared in a 3 folder system on Google Drive:

    • Data only visible to Philanthrobotics and LFC or Solve is shared in two private folders between each pairing

    • Data visible to all groups is shared in a folder between each of the three groups.

Methodology for aligning data (queries and other step-by-step guides)

Kartik employed Google Sheets to create a shared data schema by aligning applicant data from partners Lever for Change and MIT Solve. He utilized shared keyword queries to pinpoint similarities between fields, with queries referencing terms like "theory of change", "benchmark", and "progression". By placing the fields of each group in individual columns, Kartik removed non-matching categories and marked specific categories for Kevin's review to double-check if Solve’s description of the data matched LFC’s description of their data for a given field.

Collecting of Shared Fields and Cross-referencing Organizational Documents for Shared Description

A collection of descriptions was then distilled from the shared column names. This draft collection can be found here. Of note, there are even some column fields that were “near misses” or things that we found similar in each of our partners data but are not quite exact matches.

Current State of Shared Schema (with date updates)

  • Shared schema introduced - June 9, 2023

  • Data from MIT Solve Brain Challenge added to pool - July 31, 2023

  • Data from LFC API added by keyword search added to pool - July 31, 2023

  • Data from Wellcome Trust Mental Health keyword search added to pool - August 30, 2023

Related draft: Google doc

Comments
0
comment
No comments here
Why not start the discussion?