Skip to content

Step by Step Towards an Advanced Analytics and Artificial Intelligence Center of Excellence


Innovating with Data at Coocique, a credit union in Costa Rica

At ixpantia, we collaborate with your data team on initiatives and projects in data science, engineering, and infrastructure. Contact us at or Schedule a discovery call!

The Challenge

Coocique is an open savings and credit union in Costa Rica with a 55-year market presence, serving over 100,000 members from various sectors including agriculture, livestock, industry, tourism, services, and trade. The credit union has about 400 employees who operate 22 branches, managing over $457 million in assets.

As part of an internal reorganization, our business intelligence unit has evolved from addressing commercial issues to providing comprehensive support to the entire business operation. This significantly increased the volume of requirements and the speed with which we needed to process data and empower decision-makers with advanced analysis. Sometime after this change, a significant portion of our members began to feel the impacts of the COVID-19 pandemic, which demanded that we take our data-based innovation capacity to a new level to improve our services.

The Solution

Our Board of Directors outlined a strategic path for the credit union, where data-driven decisions would become increasingly important. Achieving this vision involved acquiring new tools and knowledge, as well as guiding the organization towards a data-centric culture.

The first step was a series of training sessions by ixpantia, from an introduction to R to a course on deploying data products with Docker containers. We invited both the existing business intelligence team and professionals from other areas such as risk, financial analysis, collections management, sales management, information technology, and auditing to participate. We aimed to democratize the tools and practical knowledge in data science across the organization, instead of centralizing it.

The second step was the development of a risk model in R, which determined the probability of default by borrowers, in which most of the design and implementation were executed by ixpantia. This early success, achieved in a short time, gave the credit union a model to follow for future projects.

The third step was entering into a collaborative relationship with ixpantia, where the execution responsibility now lies with the Coocique team and ixpantia stands by our side as an ally to ensure that our team has access to the knowledge and experience to execute at the necessary speed.

Ejemplo de proyecto de datos:
Plan Prospera

As the country and the world faced the pandemic, the credit union needed to respond to affected members. Within weeks of the government's sanitary measures being implemented, we received a list of 4,300 members reporting being adversely affected by the pandemic; this translated to 8,000 at-risk loans. It was clear that the entire data team's priority had to be directed toward this.

Developing an initiative aligned with Coocique's vision of social and economic development for our members with a strong sense of belonging was essential. However, profitability remained important as we needed a solid institution to serve our members. We had no time for mistakes! The Board of Directors agreed to deploy the project called Plan Prospera to seek this balance. Among the defined projects in the plan were the following questions:

  • How can we direct limited resources in the best possible way to help members during the pandemic?
  • How to prioritize support for members based on their history with the credit union and their level of being adversely affected by the pandemic?
  • What variables can we use to create an indicator to prioritize based on the member's importance and individual need?
  • How to segment clients according to their different characteristics and levels of impact to direct differentiated and precise attention strategies to the various needs of our clients?
  • What should be the main financial scenarios and key indicators to be adequately monitored so that actions are sustainable over time?

We complemented our experience and knowledge about our members with a methodical approach to find foundations in our data, thus identifying and selecting variables to create indicators. These indicators reflect the need for member assistance and the potential impact on the credit union's sustainability. Combined with this, this work gave us a framework for managerial decision-making that led to a better-informed pandemic response, both to our members and to the financial entities we work with.

Why ixpantia?

ixpantia works transparently. Minimizing dependence on them as a provider, they make their processes and actions fully accessible to us. During model development, they present weekly progress and all code and data transparently. In the end, Coocique had access to all the code for data cleaning and model training, as well as methodological documentation that facilitates maintaining and improving the model in the future.

ixpantia seeks to make its client independent. Their attitude of wanting to explain how things work and transferring knowledge and best practices to the organization is highly valuable. This fosters a learning narrative throughout the collaboration that helps promote a creative and positive attitude toward data use in the organization. Today, we do things more methodically, robustly, and effectively.

ixpantia has the knowledge and experience to successfully deliver projects. The projects we started became more experimental, conceived from a question or intuition, without being certain if we could complete the project with the available data and methodologies. Being able to count on ixpantia's team's decades of combined experience has allowed us to carry out these high-impact projects.

Benefits for Coocique

We love that we now do things scientifically. Beyond the knowledge and best practices learned during the training sessions, and having successfully deployed five models in a year, ixpantia's collaboration facilitated the automation of several accounting reports required by the state regulatory body. By automating and using Rmarkdown, we have reduced the time it takes to generate reports from weeks to a couple of days.

Additionally, ixpantia was already supporting Coocique through collaboration projects when extraordinary analysis needs relating to the COVID-19 pandemic arose. This allowed us to act proactively and quickly, creating data-based mechanisms to give proper priority to help and support our members affected by the pandemic. The combination of our internal team's knowledge with ixpantia's experience and methodologies allowed us to execute our COVID-19 response plan effectively.

Today, Coocique is building its Advanced Analytics and Artificial Intelligence Center of Excellence to provide a space where a multidisciplinary team can continue working on the best practices they have managed to implement.

This blog is maintained by the ixpantia team and a community of data-driven people, which we are happy to be a part of. Do you have an idea to publish something here? Send us a line!! We are always interested in new material and ideas. © 2019-2022 ixpantia