Cafélytics

2012: Past

In 2012 Michael Pilosov, founder of Mind the Grow, traveled to a coffee cooperative in Guatemala as a volunteer. For his volunteer project, he decided to apply his mathematical and coding abilities to analyze strategies for the co-op’s agricultural development.

Michael discovered that the present operational state of the co-op (assuming no additional acquisitions of land) depicted large dips in future harvest yields. These stemmed from the cyclic nature of the coffee tree itself.

Michael engaged with the managers to discuss possible strategies to address year-over-year variation in the yields on the horizon. It was necessary for the co-op to execute decisions that would level these production trends to ensure the long-term sustainability of the cooperative.

Methods

Situation

The coffee cooperative, established by a non-governmental organization, has helped lift dozens of farmers out of poverty by providing access to micro-loans, organizational planning, and marketing assistance. The rapid development of progress promotes yield trends susceptible to the growth patterns of trees. Unless planting timelines are properly staggered, steady yields are difficult to maintain.

Task

Analyze growth of cooperative, ensure longevity of operation by assessing growth strategies on a 50-year time horizon. Determine when to buy land, how much, and assess various strategies to mitigate risk of failing to meet sales contracts.

guate-farm-top

Approach

Develop model of cooperative-level yield dynamics, simulate growth cycles, and predict harvest outcomes.

Result

Code and report. The code was written in a language with a licensing structure, and it could not be handed off for re-use by cooperative managers. Despite efforts to make it user-friendly and well-documented, poor coding practices were followed due to the extremely short nature of the engagement.

Discussion

Although the code was created and the solutions were developed, the integration was not properly addressed. The managers of the co-op did not have the skills nor license to utilize the code, and thus, they could not implement it into their operations.

It was this lack of delivery that planted the early seeds for Mind the Grow, which Michael founded in July of 2018. Years of experience in the development of mathematical theory and implementation with production-grade open-source software provided fertile opportunities for practical application.


2019: Present

Mind the Grow has since expanded and rewritten the core software to create applications in a range of up-and-coming agricultural industries, such as Cannabis and Yerba Mate. The focus on scalability and cloud-based deployment is now front-and-center, informing early decisions in the software development process, with an emphasis on usability through accessible web-applications. We believe in supporting open source software and the communities that build and utilize its products.

As of this writing, it has been a year since Mind the Grow was founded. For our one-year anniversary, we have decided to open source the original code used for analyzing the coffee cooperative. The code has since undergone hefty revisions for readability, reproducibility, and with an eye for the future. However, it was not expanded nor functionally improved, as it was meant to serve a narrow primary purpose: to re-create the original results presented to the NGO in 2012.

Methods

Situation

Software solution did not translate to real-world application. Many poor coding practices were reflected due to inexperience (Michael only had one semester’s worth of coding experience at the time).

Task

Address problem of inability to re-use code due to licensing. Release code in hopes of finishing what was once started and ensure it is accessible by those who need it.

guate-farm-bottom

Approach

Make necessary changes to code in order to leverage a similar open-source alternative. GNU Octave was a suitable candidate. Re-structure code to conform to better functional programming principles. Leverage recent developments in cloud computing to deploy the software in such a way that anyone with an internet browser can access the predictive analytics tools built in 2012.

Result

Published a repository on Github and ensured the computational environment was reproducible using Binder.

Discussion

This code, an attempt at democratizing data-analytics, provides a sampling of the products and services that we offer at Mind The Grow.

Future

We at Mind the Grow would be happy to dedicate our time to building out requested features and making the model more suitable to the needs of those with interest in using it. If you would like to contribute to the development of the code or request new features, please consult the Contributing document on Github.

A lot of progress has been made since Michael’s initial foray into agricultural consulting. From incorporating the latest advances in mathematics into production-level software, to deploying horizontally-scalable web applications at scale, the breadth of possibilities is now endless. We invite you to take a look around at some of our refined software product offerings, and would love to hear from you with any thoughts or suggestions you may have about our work.


See Also:

Seed a Conversation.

Find out how we can grow together.