More data is not always better. Considering the local data that is most relevant to your decision is more important than big data.
Small vs Big Data
Small Data refers to observations from local fields, which includes pest pressure, planting details, crop growth stages, weather, and yield.
There is a reason that universities supply different forecasts and recommendations, (such as crop growth stage models and fertility recommendations) to address the locality of agriculture. There are real differences in environment and production practices across regions.
In the pursuit of big data, the value of an analysis could actually decrease when data that is irrelevant to your customers is included. Examples include considering crop performance on soil types you don't see and varieties that your customers don't use.
You Have the Best Data
As an agronomist, your experience and data put you in the position of offering the most “local” (i.e. best) advice for customers. And you likely do have “enough” data. Don’t buy into the myth that you need big data to learn anything from analytics.
You are working with fields across your region that experience similar soils and weather patterns. This allows you to compare apples to apples when benchmarking field performance. With your observations in hand, supplemented with some equipment and weather data, you can tease out the differences in practices and treatments between fields. Then you will better understand what drove performance without ignoring local impacts like storm damage and pest infestations.
Sometimes you need Even Smaller Data
Note that when you are honing in on a decision, even your "Small Data" can be narrowed down. You can select a specific group of fields, local soil types, or types of management practices to evaluate. This focus can give further confidence that the data set you are referencing is relevant for your decision.
Small Data and Ag Analytics
This "Small Data" collected by agronomists is extremely relevant to many recommendations. However, with emphasis on big data, the focus has been on algorithms and computers to aggregate and analyze large volumes of data, often without human input. There has been relatively less effort on tools to work with ‘Small’ data.
We intend to tackle this challenge with tools and services to process and analyze 'Small Data' and allow agronomists to zoom into the data that is most relevant for their decisions.
For more on this topic, check out this post from The Antara Group from December: https://antaraag.ca/2020/12/15/the-big-problem-with-big-data-in-agriculture/.
And here is a more technical discussion of including just the right data in a data science project: https://towardsdatascience.com/feature-selection-how-to-throw-away-95-of-your-data-and-get-95-accuracy-ad41ca016877.
To receive these posts in your email, subscribe below: