Progress 09/01/20 to 01/29/23
Outputs
Target Audience:We have engaged three Iowa fams: Mustard Seed Farms, Root to Rise farms, and SimpleLife Farms. We have prepared surveys to query about 500 organic farmers in Iowa. Before sending out the survey, we plan to conduct a focus group study of six to nineIowa small produce farmers. Changes/Problems:During our project, we realized from field experience and feedback from the farmers that they preferred a software solution instead of the proposed hardware-based solution. Based on this feedback, we repurposed our goals to build a software-based solution and have it available via an easy-to-access web portal. We continue to refer to this portal-based application as SPIDER, as originally named. We also moved away from hardware-based data input (e.g., using a finger, stylus, or keyboard) to voice input. Further, we used artificial intelligence tools to convert voice input to actionable data. What opportunities for training and professional development has the project provided?We have employed several interns, both graduate, and undergraduate. Our interns represent minority females in the STEM field. These interns helped BDiB in our research efforts and SPIDER application development. Also, we engaged with three freshmen honors students from the computer science program of Iowa State University to work on the SPIDER technology. How have the results been disseminated to communities of interest?We have demonstrated the capabilities of our SPDER application to participating farms, namely Mustard Seed Farms and Root to Rise Organic farm. We now participate in the NSF-NIFA multi-university consortium research project led by Iowa State University (AIIRA). We also are a principal participant with Translational AI Center at Iowa State (TrAC). What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
We developed a web-based application called SPIDER. SPIDER assists the farmer in gathering all of the daily farm activitiesthrough voice input. With open AI, the audio input is transformed intostructured data and saved in a protected and secure database. Only authorized and authenticated users (farmers and BDiB developers) will have access to the reports and information in the database. The farmer will eventually use this information to submit an application for organic certification.
Publications
|
Progress 09/01/20 to 08/31/21
Outputs
Target Audience:Project Director Dr. Reddy, Co-Project Director Dr. Nilakanta from Big Data in a Box, LLC (henceforth BDiB) and BDiB Ag Consultant Kenneth Beamer visited multiple farms to discuss with the farmers dealing with vegetable and fruit crops as well as row crops. The discussion revolved around the crops they grow, challenges faced by them in decision making process every year, problems related to data collection and to know what they would need to make the ideal decisions. BDiB also checked with the farmers regarding their interest in obtaining organic certification and challenges related to obtaining and maintaining the organic certification. BDiB decided to work with two small vegetable and fruit crop farms: one natural food growing famer (Mustard Seed Farm) that is interested in obtaining the organic certification for part of the farm and one farmer (Root to Rise Farm) who obtained organic certification in spring of 2020. The idea is to see how BDiB device can help the farmers in storing data towards obtaining organic certification or to collect data to renew the organic certification respectively. We also will be working with one mid-size row crop farmer (Story County Farm) to see how our device can help in collecting the data and to see if they would be interested in future organic certification. Changes/Problems:Problems Encountered During the interviews with farmers regarding their pain points in terms of meeting the USDA NOP compliance requirements, the issue about not having to type too many times to enter the data was a challenging one. We were able to resolve it by planning to use voice recording. We are currently able to record voice when the device is connected to wifi but we are trying to figure out a solution so that voice input can be recorded without wifi. The other challenges we are facing are converting the voice data to automatically get transferred to the intermediary tables and conversion of digital images into data. We hope to figure out a way for all or most of these problems. We are fine tuning conversion of the voice input into digital data. What opportunities for training and professional development has the project provided?Female under-represented minority undergarduate and graduate STEM students from Iowa State University participated in our research projects as paid interns. How have the results been disseminated to communities of interest?
Nothing Reported
What do you plan to do during the next reporting period to accomplish the goals?Successes To Date The first success was for Dr. Reddy, Dr. Nilakanta and Ag Consultant Mr Beamer be able to collect the requirements of the farmers and classify them into major and minor pain points. Major pain points are 1) inability to type too many characters while collecting the data in the field as hands get wet and/or dirty very fast, and problem in storing and 2) finding the sales receipts, input expense receipts etc. We were able to provide voice input recording and digital image capture as solutions to the above two problems was our first success. However we are still trying to fine tune the voice data input and digital image capture. We are able to record the voice without the wifi. We are now recommending the farmers to do voice recordings in the room or barn to avoid the interference by the wind. The other success we had was realization that the voice recordings done using key phrases has improved the ability to transcribe the digital voices. Tasks in Progress Possibility of IP: We request the farmers to use key phrases to record their daily activities. Because of the context sensitive key phrases in transcription, we are in a position to use them to convert them into our data fields. Since converting the transcribed data into the data fields is a time-consuming process. we are adapting the AI technology to the transcripts to extract the relevant information using our proprietary algorithm and feed it to the application that we developed. This is unique to us and we intend to claim IP on it. Once we are able to test and validate the effectiveness of this approach, we would be in a position to claim IP. The development of exit surveys will be done after we validate our technology and analysis of data from all the tasks will be analyzed towards the end of the project which would be towards the end of the year as we have received the extension of the project.
Impacts
What was accomplished under these goals?
Task1.1: Kickoff meeting, farmer input and goal setting for data collection process and testing The various steps involved in obtaining the seeds and or cuttings, field preparation, soil testing, water quality testing, seeding/planting, fertilizer/pesticide/herbicide application, harvest, produce cleanup, produce shipping/delivery and/or sale of produce were discussed for each farm. Mustard Seed Farm delivers the produce directly to their customers through volunteers, Root to Rise Farm arranges for pickup and direct sale in local farmers market while Story County Farm sells it to the local coop or other entity based on demand. Task 1.2: Collect data requirements for crop types: Two major requirements suggested by all farmers for any ideal electronic device for their needs included the following i) an alternative to entering data by reducing the necessity of using touch pad as the fingers of the farmers can be big or rough or dusty and hence hinder the typing using touch pad ii) having picture capture capability. Based on the discussion with farmers, we also identified the following main gaps for three farms. Mustard Seed Farm (Interested in organic certification): Currently have data mostly through hand written notes, need help in converting them into electronic format. Need easier way for direct electronic data entry, can use BDiB device. No formal field map and data driven crop rotation decision making process exists. Don't have information as to what exactly is needed for organic certification even though they do not apply any product that is not in the National Organic Program (NOP) substance list. Need help in identifying the best yielding fruit trees. Root to Rise Farm (Interested in renewing organic certification): Currently have data mostly through hand written notes, need help in converting them into electronic format. Google Earth based field map, can number the fields in the farm so as to make crop rotation decisions easily. May need help with electronic data entry. Need easier way for direct electronic data entry, can use BDiB device. Story County Farm (currently not interested in organic certification): Have a different device for electronic data entry. May need information about help with electronic data entry. Need to be informed about NOP. Compare data entry process in the currently used expensive device vs BDiB device Task 1:3: Design/implementation of data collection process at participating farms We have outlined the steps needed for the different farms based on their requirements and made a plan of action towards the compliance with certification needs based on USDA NOP guidelines. PI Dr. Reddy discussed with Ag Consultant Kenneth P. Beamer to develop intermediary data collection tables shown later in this section. The idea is to test the voice input being automatically converted into digital data that can automatically get transferred into these tables. Table 1. Inputs including purchase receipts, product labels of inputs used for fertility, pesticide, Irrigation management Table 2. Seed inputs including purchase receipts, enquiry from atleast three sources, planting, transplanting Table 3. Test ouputs: including lab analyses reports of water and soil Table 4. Harvest output including storage, sales and left over harvested produce Task 1.4: Farmer interview/surveys and development of support documents Summer intern and Iowa State University student Adriana Gerardo worked with the farmers at Mustard Seed Farms and Root to Rise Farm. Adriana interviewed the farmers and found about the improvements that can be brought to use the voice recordings and digital image capture. Farmers are happy with using the SPIDER tool for the possible outcome they derive but practical problems like the device going into sleep mode and unclear recording when wind is high, sweat running onto the devices etc. were pointed out. Task 2.1: Pre-test and develop data collection procedures to meet certification compliance We plan to develop data collection procedures using the Operating System that can integrate the communication and software development all on one platform. Project Director, Co-Project Director, Ag Consultant and Programmer Indira Reddy discussed and decided that the data collection will be done in different folders that are directly related to the NOP requirements. This provides an ability to easily take pictures of documents related to a given folder and store it there for later use. This alleviates the problem of losing documents by capturing the images soon so that the time stamp also can be used as the date of a given activity. We decided to develop and fine tune these data collection procedures to address the organic certification process of Iowa Department of Agriculture and Land Stewardship (IDALS) as the farms we are targeting for this work are all in Iowa. The intention is to focus on research leading to appropriate data capture for the given farm for alignment with NOP compliance. Task 2.2.: Choose certification guidelines for specific agencies The certification guidelines referred on IDALS site https://www.iowaagriculture.gov/AgDiversification /producersPage.asp will be used for developing our device. The idea is to get the device tested by farmers based in Iowa. Since each agency has their own guidelines, we eventually want to develop tools to meet the requirements for the agency chosen by the individual farm. However, we want to demonstrate the efficacy of our research using the IDALS NOP requirements as the benchmark for a transitioning farm (Mustard Seed Farm) or a continuing farm (Roots to Rise Farm). Task 2.3: Test, analyze, and validate data collection procedures The data collections procedures are still being standardized and once that is completed, we plan to develop a library of procedures for accurate record-keeping. Currently we are focused on recording and converting voice data to automatically fill in the intermediary tables that we created. Task 2.4: Identify scope for system testing and desired additional tests Since the season began in the end of April, we took extension and tested the edge computing, to determine the applicability of the SPIDER tool and software for different organic crops and produce. Task 2.5: BDiB tool and software specifications and development and economic feasibility Voice input using key phrases is helping in transcribing the digital voices. As described in the section Task 1.3, different intermediary data collection tables have been designed for the standardization of this method.
Publications
|