big data implications - terry griffin - 6
TRANSCRIPT
Big Data Implications for Agriculture
Terry Griffin, PhD, CCA
Cropping Systems Economist
Department of Agricultural Economics
@SpacePlowboy
#PrecisionAg #BigData #FarmData
#EOCC17
Eastern Ontario Crop Conference
Kemptville, ON February 14, 2017
Farmer’s Use of Precision Ag
0
25
50
75
100
1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015
Per
cent of
Far
ms
Automated Guidance Lightbar Automated Section Control Precision Soil Sampling
Yield Monitor Yield Monitor with GPS Variable Rate Fert. Variable Rate Seed
www.presentationmagazine.com
www.presentationmagazine.com
www.presentationmagazine.com
www.presentationmagazine.com
Data Analysis Service Offerings
82.0%
38.7%
19.5%
12.3%
9.2%
0% 50% 100%
Print Maps for Customers(Yield/EC/Soil Maps, etc.)
No Aggregate Data; Individual FarmData Only
Data Aggregated Among Farmers ButNot Outside the Dealership
Data Aggregated Among FarmersIncluding Those Outside the
Dealership
Do Not Help Customers With TheirFarm-Level Data
% of Respondents
2015 Base: 261 respondents
Erickson & Widmar, 2015
Managing Farm-Level Data to Assist Customers in Decision Making
Small data
Big data
Big data
Precision profitability over time
0%
10%
20%
30%
40%
50%
60%
70%
80%
2003 2004 2005 2006 2007 2008 2009 2011 2013 2015
% o
f re
spondents
off
eri
ng p
recis
ion
serv
ices
makin
g a
pro
fit
Soil Sampling
Single Variable Rate Application
Multi Variable Rate Application
Satellite Imagery
Yield Monitor Data Analysis
Total Precision Package
2015 Base: 261
Erickson & Widmar, 2015
Privacy and Security
• Privacy not a new idea
• “Security” & “privacy”
– Prevent others accessing data
– Prevent data becoming corrupted
– Prevent data loss
Source: https://www.google.com/trends/
Source: https://www.google.com/trends/
Thinking about Farm Data
• Data is intangible and irreplaceable
– “non-rival”
– “Excludable” and/or “non-excludable”
• Copies of digital data identical to original
• Anonymity does not exist with big data
• Value lies in its use, not in the possession
– Data tombs are common (and worthless)
– ‘data has no value’
My Favorite Movie
My Favorite Movie
Ownership framework
• First: what does it mean to “own” something?– Right to POSSESS*
– Right to USE
– Right to ENJOY
– Right to EXCLUDE OTHERS FROM*
– Right to TRANSFER
– Right to CONSUME or DESTROY*
• Better question: What recourse do people have with respect to misappropriated data?
Stolen from: Shannon Ferrell, Oklahoma State University
Framework for intangible property
• Where does farm data fit, if at all?
• Intellectual property
– Trademark
– Patent
– Copyright
– Trade secret
• Ellixson & Griffin, 2017 U.S. Const. Art. I, §8, c. 8
Stolen from: Shannon Ferrell, Oklahoma State University
Community Data Analysis
• Community Participation: Value to Farmer vs Network
• Value of primary use < value of secondary use
Image credit: Fox Photos/Getty Images
Data Primary Use Secondary Use
Yield monitor data Documenting yields
On-farm trials
Splitting crop shares
GxExM analyses
Soil sample data Fertilizer decisions Regional compliance
Algorithm development
Scouting Spray decisions Regional analytics
Early alerts
As-applied fertility On-farm trials
Compliance
Algorithm development
Single Field vs Community
Are value of secondary uses > primary uses?
Need 1 field Need many fields
Farm Data as Intangible Resource
• Reluctant to share data
• Ramification of relinquishing control?– Gives up bargaining power
– Fear own data used against them
Source: Shanoyan and Griffin
Valuation of Precision Ag Data• Consider the farm-level value of ‘lost data’
– Pirate holding data for ransom
– Willingness-to-pay for data security
• Spoiler: typically low compared to collection effort
• Court system likely decide value
– Rather than free market
Farm Data Valuation
0
5
10
15
20
25
30
2010 2015
Mbit
per
second
Are FCC-Defined Broadband Speed Enough?
down up
Broadband Speeds Enough?
• UAV imagery example (Buschermohle, U of Tennessee)
– 40 acre field with 17 pictures ~ 111 MB (almost 3 MB/acre)
– 92 acres with 152 pictures ~ 450MB (almost 5 MB/acre)
• Other sensor and prescription data (Shearer, tOSU)
– Spraying 0.3 MB/acre
– Planting 5.5 MB/acre
– Yield data 4.2 MB/acre
– Soil/Fertility Data 0.6 MB/acre
– Prescription files 0.01 MB/acre 0
10
20
30
2010 2015M
bit
per
second
down up
FCC Broadband Definition: 2010 to 2014
Source: http://www.broadbandmap.gov/speed
FCC Broadband Definition: 2015
Source: http://www.broadbandmap.gov/speed
Farm Data Quality
• Yield monitor data is great, but…
– Sensor calibration crucial
– Yield cleaning necessary
– Data ≠ truth
USDA ARS Yield Editor
Raw data from combine yield monitor
Default cleaning algorithm
Advanced data cleaning algorithm
Opportunities
Stolen from: Shannon Ferrell
“Future” of Farm Data
• Wireless infrastructure impacts farmland values
• Secondary uses recognized as valuable
– If yield monitor malfunctions, harvester stops for repair
• Data quality viewed as important
– Small data at the grower:farm:field level
– Big data at the community level
• Going off grid not sustainable
Acknowledgements
Dr. Shannon L. Ferrell
Associate Professor, Agricultural Law
Dept. of Agricultural Economics
Oklahoma State University
Dr. Aleksan Shanoyan
Assistant Professor
Department of Agricultural Economics
Kansas State University
Terry Griffin
Cropping Systems Economist
501.249.6360
@SpacePlowboy