Millets - specific Focus on crop choices

Following three phases needs to be followed in sequence for best results. 1. Phase one - Control moisture content / Drying in sunlight. 2. Phase two - Separation of equal-sized grain by sieving (Separate by passing through a sieve or other straining device to separate out coarser elements). 3. Phase three – Husking (The removal of covering). 4. Impeller adjustment techniques Increase/decrease the impeller inter groove distance (A long narrow furrow cut), motor RPM. The need for multiple husking and broken grains was noticed. End user products like biscuits, and cereal processing offer good opportunities for small-scale businesses because raw materials are readily available, most equipment is reasonably affordable and if the products are chosen correctly, they have a good demand and can be profitable. ● Small-scale food items like cereal processors are confronted with strong competition in the domestic and regional markets. ● To be profitable it is essential to have high-quality products, an attractive package where appropriate, and a well-managed business. The purpose of this manual is to guide the small-scale cereal miller and baker in the Nature Labs project venue to optimize their processing methods and implement GMP (Good Manufacturing Practices) and quality assurance schemes and build their technical capacity for improved market access and competitiveness. Different types of food processing can be categorized into 1) Primary processing (post-harvest operations including drying, milling, etc.). 2) Secondary processing (e.g. baking, frying, etc.) in which raw materials from primary Processing are transformed into a wide range of added value products that are attractive and add variety to the diet. We are an agriculture advisory and agriculture business. Much more than marketing products and services around the world, this means establishing deep roots locally and leveraging our multinational presence for operational advantage. ● Provide a description of the organization’s approach, methodology, and timeline for how the organization will achieve the objectives of the Terms of Reference (TOR). ● Provide a detailed description of how the management for the requested services will be implemented in regard to the TOR. Nature Labs has been associated with IBM GBS and Google cloud BigQuery Machine learning with Vertex AI for prediction. The adapted solution and timeline are proposed to meet the following requirements. AEON is the next-generation innovation service for the data journey in the hyper scaler. The technology landscape of AEON builds with data ingestion, refinement, and visualization. The SerpAPI connects to the data sources and generates the dataset as per the business scenario. The smart algorithm brings precision to data engineering. Unification is the key differentiator in AEON that reduces the effort on the data pipeline. The smart architecture framework is illustrated here. It follows the natural path toward the best possible cloud-native features. The technology pillars redefine the concept of a big lake to retaliate against any data flood. The scientific algorithm defines very specific roles for each pillar in corresponding to the business rules. AEON germinates an elegant dataset from any search engines and open datasets. The algorithm performs data cleansing and refinement with optimal sizing before the data is ingested into the data analytics engine. The design emulates distributed dataset processing for any machine learning model. To summarize the technology landscape of AEON SerpAPI for effective data collection, The python application with an algorithm for data refinement. Big Query operates for the machine learning models. Vertex Artificial Intelligence brings the finest prediction. Google data studio for reporting and visualization layers. Human-driven process of handling a large dataset in the agriculture advisory demand a large dataset size of a zettabyte (~ trillion gigabytes) and is error-prone. The inherent complexity related to data cleansing is very time-consuming and is a root cause affecting a prediction matrix decision process. This is still an unresolved and ongoing issue in global practice in Ag. Methodology of the proposal: In order to solve the said problem within the delivery timeframe, an algorithm was needed. Nature Labs has designed an algorithm: ● with the ability to handle any data sizing and data types ranging from structured to unstructured ● with high processing speed and quality in data cleansing ● with replicability in any industry ● without compromising the data integrity. Nature Labs has developed an algorithm named Applied Engine On Nature (AEON) is designed with: A bundle of machine learning models of BigQuery Machine Learning and Vertex AI The python application wraps with SerpAPI for google search and automatically generates the optimal dataset. The ML model trains on the generated optimal dataset. The AI-based algorithm is intelligent to search once, and the fetched results are read multiple times hence reducing the cloud utilization billing cost. The key features: automatic generation of optimal BigQuery as per ANSI standard creation of an external table schema with error-free SQL syntax embedded with dataset path. The proposed solution to meet the requirements: 1) Through a consultative design developed in-house digital transformation algorithm named Applied Engine On Nature (AEON) for output-1 and for output-2 the farmers producer organization will be engaged in the proposed states of Maharashtra, Himachal Pradesh, Karnataka, and Assam. 2) The Solution will be carried by field engineers at identified villages. They will be stationed and work closely with farmer’s community, Government agencies for the field work. The data ingestion process will start from the field level and outcome of AEON will be educate through the field engineers for the beneficiaries with our experiences senior leadership team for field study and field-based process. Project Execution Stage An in-house algorithm that Natue Labs developed is named Applied Engine on Nature. AEON has 3 pillars. ● BigQuery - Machine Learning. ● Vertex AI; and ● The python application is wrapped in SerpAPI. The application layer automatically generates the schema Data Definition Language irrespective of data types. This feature eliminates all the complexities in big data architecture. Aeon search once and fetched results can be visualized in multiple ways. It optimizes the cloud billing cost considerably. The project is in a steady state. Now, over 100+ practitioners are implementing it in more than 10 use cases. The comparison on the technology landscape before and after the implementation of AEON. Before AEON, multiple Bigdata engineering tools such as Hadoop, Spark, and Cloud database that is MongoDB were used. The variety of ML models consumed a lot of time & effort. AEON implementation brings unify data pipeline solution through a big lake connecting different data lakes and datasphere. Also distributing the machine learning data into different distributed workloads. A Convolutional Neural network was used agriculture images and linear regression was used for the tabular dataset. With the Vertex AI, the accuracy of prediction is improved. We achieved the solution with modern data orchestration. The algorithm simplifies the data pipeline by reducing the complexity of collecting data from multiple sources. 1) Nature Lab will work with farmers and state environment agency for collecting the information on farmers including women farmers, in the identified geographies and also determine the gaps in information availability and its usage to contribute to the overall TCP project objectives. 2) The location will be identified based on Composite Index (CI), Natural Resource Index (NRI), Integrated Livelihood Index (ILI), Dry Days, Drought Occurrences, Area under rainfed agriculture & Gender Ratio, etc. The location selection will be done in a consultative mode with NRAA. 3) The Nature Labs will work closely with existing government agencies that are already providing physical and digital crop advisories in the identified areas. The assessment and other relevant information from the local authorities and appropriate national authorities will be used to determine the interventions across the cropping cycle – pre-sowing, sowing to pre-harvest, and harvesting to post harvest. 4) The interventions covers all the following aspects ● Hyperlocal Weather forecast and alerts for the local area ● Ground water availability ● Soil typology, and soil health ● Crop advisories ● Pest and disease alerts ● Market linkages and price alerts ● Community engagement Output 1: Crop forecasting framework and model incorporating climate (weather), soil characteristics and market information developed and piloted to aid rainfed farmers to make informed decisions Activity Constraints and models identified and reviewed Inception/Kick off Activity Develop suitable models and framework Uniqueness of the best practice - Handling a large dataset in enterprise applications is inevitable. - Hence AEON's design perfectly matches the Data Growth of Agriculture enterprise. - Our solution is aligned as Enterprise Google Cloud Adoption Framework. - AEON is unique, as it brings the best possible accuracy in terms of prediction. - Also with the power of AEON, it brings the right optimal datasets for data cleansing. - We also strike the right balance to train the data, for Machine Learning Models between over fitting and underfitting. - In terms of measurable quantities, we are close to natural prediction, hence the name AEON. - AEON almost eliminated human intervention. - Can predict close to the human way of thinking. - And the results are very evident, and is already published in 3 international papers. Establish the technology base for delivering services to farmers Product awareness with farmers Training and capacity building of farmers Activity Piloting of framework Agro-advisory & risk mitigation services to farmers First season assessment Second Season assessment Policy Brief development Final project report including data interpretation Output 2: Capacities of rural extension workers enhanced to support farmers in making informed crop choices using the framework Activity Training module for rural extension workers developed Extension training manual updated Activity Rural extension workers trained for extension services The Nature Labs will deliver the overall objective of the assignment on developing & implementing forecasting/predictive techniques and tools using BigQuery machine learning and Vertex AI for multiple data points and big data analytics to aid vulnerable farmers and planners to make informed decisions on crop choices, particularly in rainfed areas. The Project dataset carry both the predictive model and a large dataset with structured and un structured with complete scientific mechanism in collecting the whether, soil, market data related to crop and types of crops which the Natural Labs experience around 25 years. Hence with the expertise with google cloud architect and the field engineer the complete solution will be delivered with in time period October 2022 to December 2023. The application design and overall workflow is illustrated below. The above workflow is perfect solution, which brings the above required objective.