Design

Intended audience: DataKind Volunteers

Introduction to the Design Stage

Once DataKind and the partner agree to move forward with designing a project together, they move into the second project stage: Design. In this stage, DataKind and the partner organization collaborate closely, starting with essential legal agreements, to creatively conceptualize and plan their project. This phase focuses on assessing data, refining project ideas, and establishing clear goals, ultimately culminating in a detailed Project Brief aimed at achieving sustainable impact as outlined below.

Human-Centered Design

Human-centered design is a critical component of the Design stage! We define “users” as the users of the system/tool developed and the people affected by the system/tool developed. It is critical to engage both end user groups in the design process. Otherwise, you’ll likely end up with a solution that is never used. Or even worse, you might have a solution that is implemented in a way that does harm. Work with your users to collaboratively write your problem statement and define your project scope.

Incorporating good design principles into scoping your data science for good projects will take us beyond the initial excitement of a cool flashy project to the real lasting change, moving the needle on tough social-sector problems. At DataKind, we believe that 90% of the data science for social good work is problem wrangling, and that’s why human-centered design is such an important component of everything we do.

A note on creativity and experimentation

We value creativity and brilliant ideas! All the best practices and ideas discussed in the Playbook come from over 10 years of real-world application, experimentation, and documenting the lessons learned. The Design stage is a great time to consider trying new things. However, in many cases, the most useful solution to the partner organization is a tried and true data science solution or model that simply needs to be adapted and applied to their use case.We want to provide only what will be most valuable to the partner organization.

Design stage components

Many of the following components can happen in any order, or in parallel with each other. There are only three that must happen at a specific time, which have been identified and bolded.

MUST DO before starting the Design stage
- Sign a mutual NDA and data license agreement
Complete throughout the Design stage in the order that works for your team and project
- Review data security requirements with the partner organization and create and implement a data management plan
- Complete a data audit, ensuring you have the correct and sufficient data to complete the proposed project
- Evaluate possible data inclusion & exclusion risks and create a mitigation strategy
- Consider who should weigh in on project ideas as advisors or subject matter experts and incorporate their feedback
- Identify communities impacted or affected by this work and ensure accountability measures are built into the project
- Assess project-specific risks and ethical implications of the end products, establishing a plan to ensure the project incorporates responsible data science practices
- Determine if the project is feasible and define technical methodology
- Refine the drafted project statements and decide on associated specific success metrics for the project
- Agree with the partner organization on a plan for how you will gather the success metrics to measure impact in the long term
- Identify organization’s software and tool cost limitations and agree on software and tools needed accordingly
- Draft a pathway to adoption and sustainability plan for the organization post-project to check that the project is realistically sustainable
- All parties decide to pursue the project and agree on clearly defined deliverables and timescales
- Complete your project brief, documenting all design work
MUST DO at the end of the Design stage
- Create and sign services agreement
- Scoper(s) check in with advisory team to review what was designed