USDA and Microsoft Announce Winners of Ag Data Innovation Challenge

Microsoft and the US Department of Agriculture (USDA) have announced the winners of their Innovation Challenge, a competition designed to task entrepreneurs with making good use of the USDA’s data.

This data was recently released through the White House Climate Data Initiative, which aims to harness climate data to increase the resilience of America’s food system. The USDA provided contestants with more than 100 years of crop and climate data through Microsoft Azure, Microsoft’s cloud computing platform.

The data sets also included information that Microsoft gleaned from sources, including NASS Quickstats API, ARMS Far Financial and Crop Production Practices, NASS CropSpace API, and NASS VegScape API.

Submissions were evaluated on four criteria: quality of idea; implementation of idea; clarity and accuracy of solution; and potential impact. The reviewing panel included a variety of industry professionals, including a farmer.

A total of $63,000 in cash and prizes was awarded to winners selected from more than 346 registrants and 33 submissions from around the world.

Farm Plenty, which enables farmers to analyze USDA data about crops grown within five kilometers of their farms to make informed decisions about their own crop choices, took the Grand Prize. It also won the Open Source Application Award and Best Visualization in Time or Space Award. The developer, George Lee of San Francisco, runs an app development startup company providing him with a built-in user base to test the app.

The Second Prize and an Open Source Application Award went to Green Pastures, which was developed by Khyati Majmudar of Mumbai, India. The dashboard interface allows a farmer to visualize production, economic, livestock, and commodity data from NASS, ERS, ARMS, and other national and local sources, including information on farmers’ markets.

What’s Local, submitted by Benjamin Wellington of Landscape Metrics in Brooklyn, New York, won third prize. This tool analyzes the resources that are required to produce agricultural outputs by using data from the Census of Agriculture in a way that allows urban population centers to connect with farmers in their area.

Other award winners and honorable mentions include:

• Honorable Mention: Open Source Application Award, and Best Student-Made Award recipient: Farm Profit Calculator, submitted by Fernando Napier and Matt Pedersen of Lincoln, Nebraska. Farmers can use this mobile phone application to compare their input costs — fertilizer, seed, fuel — to regional averages, and find financial efficiencies where their costs are above the norm.

• Honorable Mention: Croptrends, submitted by Chaiyawut Lertvichaivoravit and Ta Chiraphadhanakul of Thousand Eyes in San Francisco. This tool can be used for viewing spatial and temporal trends in crop production and yield by county for the entire United States by using NASS data.

• Popular Choice Award: VAIS, submitted by Ken Moini of Thallo Tech in Nashville, Tennessee. This tool uses NASS data for the entire United States to provide a unique approach to visualizing crowd-sourced pricing data.

• Large Organization Recognition Award: Farmed, submitted by Bryan Tower of Applied Technical Systems in Silverdale, Washington State. This tool allows farmers to view crop conditions in their area by using VegScape data from NASS combined with local weather data.

“The partnership with Microsoft was really fantastic and the creativity that they brought to the table, the willingness to work with us,” USDA deputy undersecretary Dr. Ann Bartuska, an organizer of the Challenge, told AgFunderNews. “We have some very cumbersome, clunky data sets. To be able to make them more publicly accessible was not a simple task. Microsoft helped us do that, and it shows the power of these public-private partnerships.”

AgFunderNews recently caught up with Dr. Bartuska to get the scoop on the winners, how the data was used, and what may be next for USDA’s partnership with Microsoft.

How many submissions did you receive?

We received about 33 submissions. We actually had a fairly large number of statements of interest—over 300. It drove a lot of excitement. I think we were limited in part by the time. People expressed all sorts of other ideas that they wanted to pursue. The challenge only covered four data sets. I feel like this is 0.1 percent of what we have access to if we made more USDA data available.

What were your initial impressions after reviewing the submissions?

First, how many different approaches people took to the same data sets, because there were only four. Also, they created apps with different questions in mind, and I thought that was really very informative regarding the questions someone might ask after first looking at the data. Part of what we want to do is validate those aspects. Farm Plenty, for example, gives an easy way to navigate around different types of statistical data. You are on your own farm in Iowa, but you can look at what’s going on in counties around you, or look at particular crops across the country to see if there are hot spots of production. The visualization that was done just made it so much more useful to get stuff out of the data. We produce it as tables, and that’s not always as useful in making sense of it.

Did you notice any trends among the submissions?

I think where they ended up was focusing more on production data. Trends I noticed were geographic distribution and geospatial arrangement of data.

Was one particular data set used more than the others?

It’s hard to say. They really were all used in one way or another. One submission I thought was fun is What’s Local, the Third Prize winner. There, they used the census, and it really hits an interesting sweet spot right now because we are talking a lot about urban agriculture and connecting urban populations with food sources, so What’s Local was a fun one to see.

If you did this challenge again, what might you do differently?

I would like to provide more data sets and different data sets. If we could have gotten some of our soil data, that would have been interesting. Also, pointing to weather data more deliberately. We had some drought information and being able to more deliberately connect those types of data sets to the production data would’ve been great. And a little bit more time.

Now that you’ve selected the winners, what’s next?

We are trying to assess where to go next. We met with the developers at the end of the recent snowstorm, so not everyone could join us. About six of them were able to make it. They all pointed out that although they had to limit their focus during the three months, they had for the challenge; they also saw other opportunities for connecting with other data. So, one question is can we work with them to refine what they’ve done? We’d like to test their applications to a larger community and find out what other types of data they might want to utilize. We don’t know how we are going to go forward, but we see an opportunity in the future.

USDA has created a website featuring all 33 submissions, allowing the public to interact with each app and take a peak at what the developer created. Let AFN know which one you’d vote for and why! Email [email protected]

Image credit: Yoel Ben-Avraham