Skip to main content
SearchLoginLogin or Signup

[Soft Launch] Mental Health: Data Analyses and Visualizations

Today we release our data analyses and dashboards: a key feature to the analytics we hope to offer the philanthropy community.

Published onDec 07, 2023
[Soft Launch] Mental Health: Data Analyses and Visualizations
·
key-enterThis Pub is a Supplement to

Work with Lever for Change (LFC) and MIT Solve (Solve) has allowed us to collect applicant data from a wide variety of teams that span both organizations. Previous work has allowed us to collect a shared schema between the two organizations (link here) and describe that shared schema (link here). This will allow the Philanthrobotics community to grow and give other organizations who would like to join a template to share data.

A Mental Health Example

As we thought about the data visualizations, we wanted to root it in a large issue that philanthropy is already tackling: mental health. We were able to collect many applications across LFC and Solve:

  • 450 applications

  • 6 competitions represented

  • 2018-2022 data years represented

  • 67 countries of origin represented

Data overview

Data available from these applications includes the following fields from the shared schema (the list is non-exhaustive):

  • Organization

  • Year of Competition

  • Competition / Challenge Name HQ Location

  • Limiting Factors to Success

  • Financial Sustainability Plan

  • Future Work Locations Country

  • How will you measure your progress toward each outcome?

  • Human-centeredness of your solution

  • Key Partners

  • Months to develop a pilot

  • Organization headquarters

  • Organization Location

  • Organization Name

Within the application for each organization, we found other similar fields that were not necessarily public

  • Solution Team / Key Staff

  • Feasibility Scoring

Visualizations

We developed a range of visualizations, augmented in part by the Our World in Data Mental Health catalog. We used the DALY (disability-adjusted life years) values for mental health and prevalence numbers across various geographies to get an idea of how many dollars/projects/calls are following various mental health ailments relative to their severity.

The following visualizations do not represent all of the visualizations we have to-date, but the best overlap of data between organizations that we have. More visualizations will become available on request or as we obtain more data.


Mental Health Project Dollars per DALY

Schema (first 3 rows shown):

HQ Location

Number of Projects

Total Project Monies

Project Dollars per DALY

Greece

2

300000

326.9831035

United States

174

15842100986

25081892.82

Canada

13

1172000

2104.500064

Mexico

5

200000

379.7634681

Using the proposed budgets of each project as the intended dollar amounts that follow each mental health ailment, we found the total DALYs for all mental health ailments researched by Our World in Data, then used that as a denominator on our aggregated budgets for each organization’s headquarter location. In the visualization, dark blue and purple countries have many dollars chasing few DALYs (relatively) whereas lighter countries have the inverse.


Percentage Breakdown: Mental Health Project Budgets by Organization Type

Schema used (first 3 rows shown):

Org Type

Number of Projects

Percent of Total Project Monies

Average Employee Count

For-Profit

43

0.05268801753

54

Non-Profit

27

0.02422117678

37

Not Registered as Any Organization

40

0.005069784419

34

We were curious to answer the questions: who is proposing the largest budgets in mental health? Do they have small (start-up) like teams? It turns out that over 60% of the budgets in the mental health calls we looked at were represented by NGO classified organizations (small red triangle in lower right hand corner, large blue triangle in upper right hand corner), and their team sizes (represented by the size of the triangle) spanned the limits of the dataset. NGOs are well represented in mental health projects, and tend to be both small and large, with larger teams proposing larger budgets. In contrast, for-profit organizations are not well represented, making up less than 10% of all applicants, while still representing small, nimble teams and budgets.


More Data Cleansing In Progress: Impacted Population Breakdown by Number of Projects, Monies (i.e., Budget) per Project, and Percent of Total (Budgets) per Project

Schema (first 3 rows shown):

Populations that will benefit from your solution

Number of Projects

Total Project Monies

Money per Project

Percent of Total Project Monies

Infants / toddlers (0-2 yrs.), Children (3-9 yrs.), Caregivers

3

1447976942

482658980.7

3.542290443

Infants / toddlers (0-2 yrs.), Children (3-9 yrs.), Parents

3

2424600132

808200044

5.931474201

Children (3-9 yrs.), Families

1

889048116

889048116

2.174942538

We wanted to get a good sense of where proposed monies (i.e., budgets) were going towards impacted populations in mental health. There is still some data cleansing to do here, as each organization has specific labels for impacted populations, but we have found that many projects propose impacted populations of infants, toddlers, and their parents overall. As we consolidate this data more, we can get a better percentage, but > 10% are focused on the family unit as the impacted population.


Research as a Main Stage of Mental Health Projects, Scaling Organizations with Highest Budgets

Schema used (first 3 rows shown):

Project/Solution Stage

Number of Projects

Total Project Monies

Money per Project

Percent of Total Project Monies

Research

282

14008288

49674.78014

0.03426948542

Early

17598371

399962.9773

0.04305216443

Growth

13

6938235

533710.3846

0.01697350477

Highlighted in this visualization is the pieces of the pie that each project is “staged” in. As we continue to work with organizations, we will hone these into nicer categories, but we have found that, when asking the question: “what types of projects make up mental health philanthropy?” research projects take up a majority, but scaling projects tend to take up the most budgetary dollars in total. Meanwhile, new partnerships have the highest dollar amounts per project.


Schema used (first 3 rows shown):

Year

Number of Projects

Log(Projects)

Total Project Monies

Log(Sum Money)

Money per Project

Percent of Total Project Monies

2018

339

2.530199698

38544894

7.585966856

113701.7522

0.09494637659

2019

10

1

4790803172

9.680408328

479080317.2

11.80102875

Over time, how has the number of projects per year changed relative to the project monies (i.e., budgets) proposed in that year? We found that both of these numbers stayed fairly stable over our dataset, but there are some particularities in that Solve’s data mainly makes up 2018 as a year. Moreover, average money per project surpassed $400,000 in 2022 and has remained fairly stable dating back to 2019. This graph is on the log scale. The log fields are calculated after dataset aggregation.

Future Directions

We will continue to take requests from organizations we work with (particularly from Solve and LFC) as well as ask for various fields as our shared schema grows. However, we have our sights on improving the aforementioned visualizations in a few ways:

  • Using total award amount rather than budget dollars as a more accurate way of tracking dollars in the ecosystem.

  • Mapping out the partnerships of each organization after de-duplications onto a acyclic graph to see geographical and organization relationship.

  • Reducing categorical groups in organization type, stage, and impact population to reduce text in various visualizations.

Connections
1 of 3
Comments
0
comment
No comments here
Why not start the discussion?