GOAT: Hack @ USDA ARS (Apr. 4 & 5). Call for Applications!

Good GOATS!

Another mini-GOAT event is upon us. We’re organizing a hackathon at the USDA Agricultural Research Service in Beltsville, MD on April 4 & 5, 2019.

What this all about? More information here.

Please remember: you MUST fill out the application form to receive an invitation. Due to logistics, we’ll need to get an official headcount prior to the event. Applications close March 15, 2019.

This thread is for you to post your ideas about what you’d like to work on together. Some examples here.

You can either:

  1. Post a concept you’re working on, OR
  2. Propose a concept related to the hackathon mission.

Onward, friends!

I’ll start :slight_smile:

Cover crops are a multi functional tool utilized by farmers to achieve a variety of goals ranging from improving soil health to enhancing water availability. A group of researchers (at ARS, NC State, University of Georgia, among other institutions) are working with farmers to assess how soils, climate, and management interact to affect cover crop performance (quality and quantity) and subsequent crop yield and yield stability under current and future climate scenarios. They are monitoring soil moisture, cover crop quantity and quality, crop performance, and farmer management history. In the CROWN project (Cover Crops: Real-time Observation of Water and Nutrients), they have collected data on 80 farms from Pennsylvania to Florida. Many tools and services exist for researchers to manage experimental data to inform future research, the development of farmer-facing tools, and to gain insights into the complexity of sustainable agricultural systems. Yet, the crux of their challenge is the lack of a coherent data pipeline that allows each of the stakeholders (farmers, researchers, extension agents etc.) to have easy access to data both current and historical. Example projects that fall out of this work include:

  • Data harmonization across 3 data streams: soil moisture data collected via in-field sensors, sample data manually collected by researchers, and lab data as a result of in-lab analysis.
  • Developing a database, or perhaps a Research-Oriented FarmOS interface, to aggregate these data streams, including an API for downstream stakeholders, for example, a data visualization tool aimed at providing farmers with real time feedback on the data collected on their farms.
  • Developing a suite of basic data collection tools based on, for example, the OurSci data collection platform, to be used by the researchers and technicians involved in the CROWN project.
  • Developing a minimal dashboard to communicate database and data flow health and status.

Pipe up if you’re already working with me on this, or would like to :slight_smile:

1 Like

We want to work on the ontology piece. We have a large, wide ranging set of partners implementing similar (but not the same!) on farm and research level field surveys (observational and hypothesis driven), and we’re hoping to make data comparison across this group as consistent as possible.

@DanT is leading the effort and we’ve been working with @julietnpn and @sudokita to identify appropriate ontologies (or taxonomies maybe?) for questions (like ‘did you till’, ‘did you add an amendment’) and answers (like ‘disked, 6’’’ or ‘compost’ or whatever), with enough context to help the data collection service present an appropriate set of results back to the service to not overwhelm users.

Our case is almost identical (yay!) to USDA’s case, and we’d love to collaborate to further expand utility and cross comparability of data.

So we’ll bring our draft ontology (which we literally need to implement this year, so year… sorry :slight_smile: and would love to use it as as a starting point of discussion, ideally with feedback that we can integrate on-site.

3 Likes

Hey y’all, just found out about this upcoming event, which looks very interesting and relevant to my own schemings. Not sure I’ll be able to make it, but wanted to share some ideas I’ve been working on, and gauge potential interest from others.

Some background info about me: After 20 years of front line food & ag work early in my life, I’ve been at Cornell’s Mann Library for another 20, engaged in a variety of outreach related activities, including support for a geospatial data repository (cugir.library.cornell.edu) and several other digital resources. I’ve also worked on outside projects over the years as a consultant and volunteer, including an experimental “Northeast Food Knowledge Ecosystem” initiative with @dornawcox, looking at ways we could more effectively link the data and information resources of Northeast food and ag groups together. Partly in response to the lessons learned from that and other projects, more recently I’ve been looking at how Land Grant institutions and libraries like mine can help do some of the heavy lifting needed to support such networked knowledge systems. That includes a yearlong “Land Grant Informatics” fellowship co-sponsored by the eXtension Foundation. I usually point people to this blog post as an introduction to that wide-ranging work (the extension site is currently having issues properly displaying formatted text and images, so referring to IA for now). I presented some of the ideas from this fellowship at the 2017 IC-FOODS conference, hosted by @mateolan and focused on creating an “internet of food”.

I’m now working on ideas for what I hope will eventually be a federated network of “FAIR food data hubs”, making ag and food related data and information resources more findable, accessible, interoperable and reusable. I’ve pasted some text from a draft one pager describing that below. My more immediate plan is to develop a proof of concept CKAN hub here in New York State initially, which will leverage its metadata harvesting capabilities to eventually support a federated network. I’d love to work with others on this and future iterations…

Cheers!

FAIR Food Resource Network Proposal (DRAFT)

Jeff Piestrak, jmp36@cornell.edu

There is increasing awareness that local and regional food systems represent an area of immense opportunity for community and regional development. The 2017 report from the Federal Reserve System and USDA, Harvesting Opportunity: The Power of Regional Food System Investments to Transform Communities, highlights growing mainstream recognition of this potential. It states:

Development of regional food systems not only contribute direct economic benefits to the community, but can also open the door for improved access to healthy food and other positive outcomes that could result in improved community health and a more productive workforce.

A key element of success in this work are asset-based community development strategies which help families, farms, communities and organizations identify, build, leverage and retain multiple forms of health and wealth, equitably, without diminishing future prospects. That includes individual, social, intellectual, financial, natural, cultural, built, and political “community capitals”.

Yet as our and others research has revealed, food systems researchers, learners, practitioners and stakeholders often face significant barriers in efficiently locating and effectively leveraging high quality data and information assets (intellectual capitals) in support of such food systems development work. Many are accessing information from an increasingly distributed array of sources, and may be overwhelmed with the diversity, complexity or reliability of those. Some are also creating, collecting, aggregating and transforming these for their own use with varying degrees of success and transparency, while duplicating efforts and potentially introducing errors.

In response to these challenges and the tremendous opportunities modern information and communications technology (ICT) and expertise offer in addressing them, we propose developing a publicly accessible FAIR Food Resource Network . This networked, socio-technical “open stack”, will link and leverage existing resources and expertise in support of locally grounded, globally connected agrifood systems learning and innovation , including “Farmer Research Networks”.

This informatics work will include engaging the resources and expertise of the existing network of Land Grant institutions, including LG universities & libraries, Ag Experiment Stations, and Cooperative Extension Systems, as well as community, state, federal and global partners (like GODAN). A federated network of FAIR Food Resource Hubs will make a variety of currently disparate agrifood systems related data and information more readily and freely Findable, Accessible, Interoperable and Repurposable (FAIR). Recognizing that the success of this distributed sociotechnical network will required development of social as well as technical capacities, considerable attention will be given to supporting “effective use” and co-design of this continuously evolving system, as well as the necessary governance structures.

3 Likes

@JeffPiestrak well I really hope you go!!!

Hey everyone!

So I don’t know if it’s a project in and of itself, but I’m always game to spin up a UI (preferably web-based) for any other project that could use one.

I’d also really love the chance to work with a group on the challenges of getting good data capture on working farms, and specifically what UI strategies can be employed to solve for those problems. In other words, how do we design our UI’s to capture farm data that would otherwise fall through the cracks? This is always the first objection I get from farmers when I ask them what they do with their data: “I don’t have time to spend working on spreadsheets,” or “How the #$%^& am I gonna enter that on my phone when my fingers are caked in mud and sun’s too bright to see the screen?”

I’m not sure if there will be any/many frontend devs or UX designers there, but it would be interesting to discuss this issue with anyone who has a stake in improving data inputs in the field.

1 Like

Yes! As an in-field Ag researcher with often mud (/manure) caked hands and lots of ‘sun too bright’ / ‘wind too windy’ / ‘rainy drizzle means I can’t use screen swiping easily’ moments – UI / data entry for Ag data would be really great to discuss!

1 Like

Hello!

My project is perhaps too broad to be of merit for this hackathon, but, roughly:

I am developing methods for the physical design of agricultural properties (farms and their buildings, roads, waterways, hedgerows, riparian corridors, etc.) that are informed by publically available environmental data (topography, soil surveys, hydrology, etc.) and stakeholder-documented observations. The ability to use environmental-data in the generation of discrete architectural and landscape form (earthworks, planting plans, maintenence regimes, building facades) is made accessible to designers working in 3D modelling environments through visual scripting tools (Grasshopper for Rhino3D, Dynamo for Autodesk Revit) and built-in Python scripting environments. The informed use of that data in a design-workflow requires a data-literacy that few designers or farmers have the resources or time to develop. I am already working on these topics through workshops I teach at the University of Virginia’s School of Architecture, but am hoping to (1) sharpen these concepts through contact with agronomists and other designers and farmers, and (2) make the content of these workshops publically available by producing content for the web and tailoring these methods to open-source/inexpensive tools (software, sensing hardware).

Embedded somewhere in here is an interest in how the tools of designers (3D modeling, construction documents), farmers (spreadsheets at our farm), and environmental-scientists/agronomists can offer better interoperability. This arises in part from my frustration with the misuse of environmental data in the field of design. I don’t know much about this yet, but y’all seem like the people to talk to.

My coding and data-skills are negligible (but growing), but I can offer the perspective and skills of a landscape architect and someone doing daily work on a farm.

A couple relevant things about myself:

I am a farm apprentice, designer, and graduate-student in landscape architecture (currently on-leave from the University of Virginia as I pursue farming for the season).

I have recently started working at Bellair Farm, an 850 acre property 10 miles south of Charlottesville, Virginia, which grows vegetables on 35–40 acres (CSA, market sales, and wholesale), raises broiler chickens and laying hens, and does a small amount of hog-raising. Our vegetables are USDA certified organic. The property is also home to a sheep dairy and a pastured beef operation. I also do freelance landscape architectural design and consulting, and am currently working with a small firm to establish new working methods as they adopt building-information-modelling into their workflows.

I haven’t participated in the GOAT community before, but have been following some of y’all’s discussions and am super-stoked by the attitudes, ideas, and pursuits.

-KJ

KJ maybe @sudokita we should have a short time where folks who can’t physically come can present a short bit about what they do and that’ll give folks the ability to ask questions and follow up with collaborations offline. Low effort for potentially high return .

KJ feels like a good fit and worth a discussion!

Full name
Shaun Bluethenthal

Affiliation
Cornell Small Farms Program

Which of these best described you?

  • Researcher
  • Applied (crop consultant, farmer)

Briefly describe your skill set
Agronomist. Ag Extension Specialist. Production-Landscape Designer.

Project Idea:
Planning and Visualization Application to Manage Farm Water

Problem Statement: Considerations for how water interacts with the agricultural landscape is paramount and the absence of design considerations for water will undoubtedly have a negative impact on farm profitability and threaten operational resilience. Solutions exist for designing away this potential burden, however, without the available time, financial resources and/ or the innate capacity for academic learning, these solutions, to the degree of comprehension that is necessary to effectively implement them, exist behind an intellectual curtain of exclusivity - the unique skill sets needed to draft tolerance-specific geometries onto unambiguous terrain features, defined as two-dimensional elevation and relief lines, are not common skillsets ubiquitous to all farms.
If this skillset could be outsourced to computational software it would allow farmers and land stewards to conceptually visualize actionable planning solutions for water management.

Project Challenge:
Given contour data (centimeter-level) and fixed design parameters*, could we develop a software program that could:

  1. identify key existing topographic features (e.g. Main ridgeline, primary ridgelines/ valleys, secondary valley, land units)
  2. Locate the desired positions for the introduced design geometries (e.g. Keypoints, Keyline cultivation patterning)
  3. Generate a conceptual map/model with the combined (existing and introduced) information

*Parameters derived using Keyline Geometry

What will you bring with you?
Agricultural expertice, example diagrams, maps, drafts, pictures, supporting video, the ability to draft the designs traditionally (pencil/ paper), and of course, positive vibes.

Anything else you think we should know?
If this project gains traction there is potential research funding to continue development beyond the Hack event.

Hi GOAT:Hackers: I wanted to drop folks a quick self-intro. I’m Peter Burkholder with 18F/GSA and I’ll be at the hackathon to provide technical support for anyone who wants to host their prototype on cloud.gov.

cloud.gov (https://cloud.gov) is a Platform-as-a-Service offering that, like Heroku, provides self-service access to application servers, databases, storage, and other backing services needed for running web applications or providing a back-end to mobile applications. For the hackathon we’re providing a “prototyping” organization for participants to use.

If you think this would be of use to you, I’d suggest you do the following ahead of time:

And that’s about it. I’ll be in the riot.im room Wednesday if folks have questions ahead of time. If you want to see the code and examples of how agencies are running open-source projects on cloud.gov, check out: https://cloud.gov/docs/apps/frameworks/#customer-example-applications

Cheers, Peter