pax.green

Tropical cloud forests in the Andes mountains are one of the planet’s hotspots of endemic biodiversity. Their ecosystem services are many: drinking water, building materials, food, carbon sequestration. Tragically, they are being threatened by climate change and illegal farming and mining.

 To save these forests, we do the following: (1) using Artificial Intelligence, we have developed a system to monitor tree cover loss and avian biodiversity; (2) to maintain this system, we are training young Ecuadorians in data analysis and computational modeling; and (3) we are thereby preparing these students for non-extractive, knowledge-based careers.

 We have codified our algorithms in an application in R, the public-domain statistical software. This application allows users to block-chain their data and model results. This enables users all over the world to track tree cover loss and avian biodiversity and to share their results in a decentralized but secure way, preserving data integrity.

Facing profound poverty, people in Ecuador’s northern Andes mountains are destroying the tropical cloud forests around them, illegally mining  gold. In doing so, they destroy the long-term services provided by this rich ecosystem: drinking water, endemic biodiversity, ecotourism, and carbon sequestration.

 Results from our AI model indicate that during the past year of the Covid pandemic, tree cover loss increased more than 20 times the rate in the preceding year. This mirrors deforestation rates throughout South America. Record gold prices, job loss, poverty, and the government’s focus on Covid-related activities have created a perfect storm for illegal mining and forest loss.

 Loss of cloud forests in our project area has profound implications for global climate change and biodiversity, triggering a negative, downward spiral of ecological disaster. The 2019 wildfires in the Amazon were themselves caused by deforestation in moist tropical forests. Habitat loss links directly to biodiversity loss. Among all regions in the world, South America has the highest number of climate-induced species extinctions.

 Ecuador’s cloud forests can be said to be yet another victim of Covid; it is particularly tragic that this victim contributes so much to the world’s struggle to fight climate change and maintain biodiversity.

Because we view our problem as poverty-induced environmental destruction, we propose a solution integrating nature, equity, and economics.

 NATURE: we have built models using Artificial Intelligence to monitor deforestation and biodiversity. With public-domain data and software, our models identify at-risk ecological hotspots in northern Ecuador at a spatial resolution of 0.25 miles and with >95% accuracy.

EQUITY: we use this technology as an opportunity to train Ecuadorians from at-risk families (some from households making $2.00 a day) in data analysis. AI will dominate the future, and if the world devolves into a “digital divide”, we are ensuring our students end up on the right side of this divide.

 ECONOMICS: Our students are learning Python, R, PHP, SQL, as well as database structures, data analysis, statistics, data visualization, and AI. They will produce data products and services that equal or surpass that of their Northern Hemisphere analysts … at a fraction of the cost, given the median salary in Ecuador is $5K a year.

 Forty years ago, South American activists called for transferring the “means of production” from landowners to campesinos. Today, our South American students’ means of production are their laptops, bandwidth, work-ethic, and their brilliant, ambitious minds.

We are primarily focused on preparing Ecuadorian university students for high-tech careers in data analysis and AI. All of them are the first in their families to attend university and their presence there is a testament to their intelligence, resilience, work-ethic, and ambition; they are survivors of poverty, class bias, and domestic abuse.

Three of them graduated from a Jesuit high school in Quito. Together—using a combination of R, PHP, SQL, and HTML—we built an information system in the cloud to integrate the data from the various offices of their school. With these data, we are analyzing the socio-demographic factors of school families that influence academic performance. We developed risk indicators to identify students in greatest need of intervention. Working on this project has given then hands-on experience in building a data system.

Five other students are products of a school in Mindo, a pueblo in the northern Andes surrounded by cloud forests. We are building another information system to analyze the data from their school. Eventually, we will be working together to refine AI models that monitor forest and biodiversity loss near Mindo. These models are transferrable and scalable to other forests in South America.

 Given the importance of these forests in combating climate change, the indirect beneficiaries of our project is the rest of the world. Already, pax.green has been collaborating with NGOs in Ecuador to provide them with the locations of areas that are losing tree cover at rates exceeding 25% per area per year.

Our solution aligns with two challenges: DIGITAL INCLUSION and RESILIENT ECOSYSTEMS.

Within the latter, we have developed and are refining an AI system to monitor deforestation and plant and avian biodiversity in Andean tropical forests; creating a scalable and replicable bistro/hostel/campus for tourists to learn about  this system; and working with local NGOs to combat illegal mining  by providing locations of deforestation hotspots.

Within the former, we are training local students from poor families in data management, analysis, and visualization and are seeking clients for their skills.

We are piloting our project in Mindo, a pueblo in the northern Andes mountains surrounded by cloud forests. This is our base to test our business model. 

We are teaching students who are now at university, and who graduated from a local school that primarily focuses on working with students from families at risk because of poverty, domestic abuse, single-parent homes, and refugee status. 

We have already done considerable work here in developing and verifying our AI model to monitor habitat and biodiversity loss. Currently, we are in the process of drafting a paper on our work for peer review and submission to the Journal for Environmental Informatics. In one year, we hope to finish construction of a bistro/Airbnb/campus that will showcase sustainable ways of living and working.

Below is a map showing the deforestation hotspots (in red) in the cloud forests of Northern Ecuador. The results are based on predictions from our AI model classifying the land use and land cover classes of each pixel (with 10 meter spatial resolution) of a satellite raster.

Initially, pax.green focused on using the latest technologies for big data, computational modeling, and artificial intelligence to track the state of the environment. In this, we have been successful, achieving more than 95% accuracy in predicting land use and land cover classes. Our algorithms codify a workflow in an R package that enable users to convert satellite images into datasets and datasets into maps that identify at-risk, ecological hotspots.

 We have been focusing on environmental goods, an externality that, by definition, lacks a market. But the skills needed to produce our technologies are the same skills private companies need to navigate the digital future. Our second source of innovation lies in our identifying, training and working with brilliant and ambitious young Ecuadorians from at-risk families to develop these technologies. Our biggest challenge lies in connecting these students with clients from the Northern hemisphere seeking big-data analysts for remote, cloud-based work.

 Our third innovation is our plan to create a physical space in the cloud forest, in Mindo, a pueblo that has staked its future on ecotourism. This space is part bistro, Airbnb, and campus. It will be a place for our students to continue to learn. But it will also be a place for tourists from all over the world to learn about our ecosystem and about the technology we use to monitor it. We see these tourists as collaborators; we will use AI to analyze the photos and videos they take and add to the database we are building.

Our project area covers 65 square miles of forested area in northern Ecuador. But the technology is easily scalable and transferrable to all forests in Southern America and the rest of the world. We are currently working with about 12 students to train them in all aspects of data analysis. Before expanding the number of students, we hope we can place existing students within jobs that will pay them well to apply their data skills. If successful, we foresee doubling the student base in five years. Our solution also serves the needs of the Mindo’s 4,500 citizens. This pueblo depends heavily on ecotourism and its people understand the link between habitat and biodiversity preservation and Mindo’s economy.

Our single most important indicator of success is the number of students we can place in positions that pay well and that provide them an opportunity to work in jobs where they can use their creative talents. Already, two of our students have leveraged their training to pursue graduate degrees in the U.S. and Argentina.

To be sustainable, we also will track the revenue we will generate in our two businesses: data analytics and ecotourism.

 Thirdly, using images from Europe’s Sentinel satellites and from Cornell University’s ebird database (both of which are being continuously updated), we can track numbers of trees lost and number and distribution of bird species in the cloud forests surrounding our center.

Four full-time staff. Three focus on ecotourism; the project lead focuses on data analysis training and developing AI models.

 Currently, about 12 students.

For the ecotourism component of our solution, the team is well placed because all of the team members are from Mindo, the pueblo where we are building our campus; a few of them have worked in the hospitality business for the past seven years.

For the data analytics component of the solution, the project lead has worked for the U.S. government on computational models for 20 years. He has spent these years developing models and training staff on all aspects of data management and analysis.

By the end of 2021, pax.green will have created two R packages—one to monitor tree cover loss and the other to monitor biodiversity. These are primarily concerns of governments, multilateral organizations, and foundations. As we build our capacity and credibility, we anticipate we will first pursue grants from these institutions.

Pax.green is being led by collaborative partners. The ecotourism solution is headed by an Ecuadorian with seven years’ experience in the hospitality business. He has worked for foreigners as a manager for their businesses; this will be his first opportunity to co-own and run his own business.

 The data analytics solution is being headed by an immigrant from the Philippines who lived and worked in the U.S. for 40 years. As the business moves forward, we anticipate creating a board with representation from our students and citizens of Mindo, the pueblo in Ecuador which serves as our home base. They are a mix of ethnicities and sexual orientation.

We hope to learn how to find global clients for our business in providing consulting services for big data analysis. We need to learn how to market and conduct outreach to find these clients.

We believe we understand the ecotourism business. But to create resiliency, we need to also build an online business where we can provide our data analysis capabilities to a global customer base. We have the technical expertise and experience in big-data analysis. We need to increase our ability to market these capabilities.

We welcome any and all assistance in building a global client base online for data services which we can provide remotely.

We are training students to conduct data analysis using the tools of big data management and artificial intelligence. These students are young Ecuadorians from families at risk by virtue of poverty, family structure, refugee status, and domestic abuse. Some of these students are from households eking by on $2 a day.

We are teaching them the skills and providing them the opportunity for a future in AI and data analytics, a future where they can remain in South America while pursuing clients in cybespace.

We are setting up an ecotourism business which is based in the Andes cloud forest. This business is part bistro/airbnb/campus--a WeWork in the forest where digital nomads from all over the world can convene. The place has solar panels, compost toilets, and an on-site permaculture farm: an example of sustainable living and working.

From this site, we are building AI models to monitor tree cover loss and biodiversity in the tropical Andes forests, a premiere are for carbon sequestration and biodiversity.

We will use any cash prize to build our student base, such as paying students a stipend, buying them laptops, and paying for certification as data specialists.

We are setting up an ecotourism business which is based in the Andes cloud forest. This business is part bistro/airbnb/campus--a WeWork in the forest where digital nomads from all over the world can convene. The place has solar panels, compost toilets, and an on-site permaculture farm: an example of sustainable living and working.

From this site, we are building AI models to monitor tree cover loss and biodiversity in the tropical Andes forests, a premiere are for carbon sequestration and biodiversity.

We will use any cash prize to build our student base, such as paying students a stipend, buying them laptops, and paying for certification as data specialists.

We are setting up an ecotourism business which is based in the Andes cloud forest. This business is part bistro/airbnb/campus--a WeWork in the forest where digital nomads from all over the world can convene. The place has solar panels, compost toilets, and an on-site permaculture farm: an example of sustainable living and working.

From this site, we are building AI models to monitor tree cover loss and biodiversity in the tropical Andes forests, a premiere are for carbon sequestration and biodiversity.

We are codifying our workflow within R packages. Part of the packages is to block-chain environmental data so that the integrity of the data and their provenance is maintained. This allows monitoring data to be decentralized and shared, but with data integrity preserved. The more data we can share, the greater will be the accuracy in predicting biodiversity and habitat loss.

We will use any cash prize to build our student base, such as paying students a stipend, buying them laptops, and paying for certification as data specialists.