The humanitarian sector provides a unique opportunity to scale advanced approaches. Having worked as a researcher and as a humanitarian program manager there are hurdles in both organizational structures to do this.
Research institutions are motivated to be innovative and create new models but there is little incentive to scale them. Often researchers see this as the job of humanitarian organizations. An advanced approach is made, and then the model is hosted on GitHub or their public facing web portal, or handed to the local institutions, but no one else knows how to run it.
Likewise, humanitarian organizations may create an advanced approach at a country program, but there is seldom time and incentive to create operational documents that are relevant to other country programs.
Centralizing Innovation for Organizational Scale
One way that Catholic Relief Services has overcome this gap it to have a centralized support team in data analytics. Mapping and data dashboard requests come to our team from across multiple projects and country programs. This provides a unique opportunity to understand the country program’s problem and propose alternative solutions with advanced analytics that can scale across the organization.
Once a new advanced analytics model is made, we create communication resources to build awareness around the approach. Some of the lessons we have learned so far are:
- Identify common business needs to make it easier to scale.
- Set up a centralized infrastructure where models are hosted and everyone has access.
- Esri’s workflow (previously known as model builder) is a great place to start modeling.
- Give team members an annual goal to create new models and train teams on how to run it.
- Make communication materials in common languages and share them across the agency.
Let’s Work Together as ICT4D Practitioners
I would like to share our work across the humanitarian sector and benefit from the comments and models of others. We have common advanced analytics business needs across the humanitarian sector that we all could mutually benefit from.
Our approach started when we couldn’t afford the 30,000 USD price tag quoted by an external provider. So, we created automated models on our own and built-up confidence in our team. This was a fantastic start to our data analytics journey, and I would love a place to convene where we can learn from and build off each other’s work.
How Can We Create a Data Analytics Group?
Please write suggestions in the comments where a humanitarian data analytics group should be formed and how we can share the advanced analytics models we develop.
By Kathryn Clifton and Paul Wiedmaier, Catholic Relief Services
Totally agree that we need to work together across the industry to help solve these common challenges together – especially when it comes to making evidence based decisions and investing in new technologies. Only by collaborating can we become a sector of modern digital NGOs! At Save the Children we have relatively new Digital Programming and Data & Analytics project teams, which have been set up in the last year to help solve these exact same problems, and we’d definitely be interested in exploring the opportunity to collaborate with peers to share knowledge and pool resources.
Members willing to take part in the development of the model can agree on meetings aimed at the development
I totally agree with this suggestion; one big challenge is organizations struggling to start from scrush something that is already up and running in other organization. In my organization we leverage on the DHIS2 platform for automates. It would be good to share and learn from each other.
How Can We Create a Data Analytics Group?
One way we can create it is
1 internet:
Internet helps in connecting the world together,
2 media house, social media platform
Or we can Create Apps like Facebook or like Instagram, but this very one will be design just for Data Analysis
3 training and purposes
That’s where the man power comes in
4 creating campaigns and awareness to spread the information to the rural areas,
You could form a subgroup in the Artificial Intelligence and Business Analytics (AIBA) Group in LinkedIn.
You haven’t really explained here what you mean by an “automated analytics model,” so I’m not sure what that means exactly. But if an organization the size of CRS can’t find $30k of value from improved data systems then it’s hard to see why you would have a data analytics team at all, which costs much more than that. By building tools in ESRI, you pretty much ensure that low-resource country governments cannot take up the tools and scale them, due to the high per-seat and data costs.
Dear John, No programs can’t spend 30K on most occasions without planning for it. We still have budget line items that we have to follow and plan for in advance. We bill back our services to programs. ESRI like other software vendors provides us good packages and support services. We have transitioned our mapping to other platforms when working with country governments on many occasions. it is seldom that ANY platform is aligned (mapping, bi solutions, digital data collection). As a result we focus on the problem/solution and how we can transition it. I did not get into our work on machine learning but we use Azure for Microsoft. Given the size of our organization this makes sense, however other institutes run algorithms on their local drive. We try to focus on meeting people where they are at vs. matching our large organizational scale and IT setup. Thanks for reading the article.
What is Automated Analytics
Automated data analytics is the practice of using computer systems and processes to perform analytical tasks with little or no human intervention.
Automation can streamline data preparation tasks. Tools like the visual programming platform KNIME can automatically label data, train and validate models, and iterate study runs to optimize parameters.
Creating & using prototypes to create much more of a safe & healthy society is what drives our innovative technology we want to see placed in socio-ecosystems. We want to help in the provision of a healthy planet now. Next step is to obtain support & funding to save lives, our planet. Tell is how to get the funding please. Thanks
We can create data analytics group with what we already have, social media, from here on we can try to access funding that will boost automated analysis.
True
Thanks for the article, Kathryn and Paul. I agree with the general sentiment of increased collaboration and building on each other’s work. In that spirit, I didn’t really understand from the article the goal of models CRS has developed. What questions are/were you trying to answer with them. I imagine you’ll want collaborators looking at similar questions.
Eric, We have models on how to optimize program activity locations so that participants walk less, where to locate shelters (multi-criteria), Tracking which households have been visited and missed, knowing what profiles have been missed by programs to improve our outreach, among others.
Hi –
You may already join this group of R users in the Humanitarian sector – https://humanitarian-user-group.github.io/ – 466 participants in the linked skype group – https://join.skype.com/qYBKC5q3wKp4
Best,
Edouard