Plot Risk Model
Feedback
What is Plot Risk Model?
It’s AI/ML Models for monitoring and providing advisories and warnings early on.
🕥
What was the Objective?
Enable the model feedback collection to ensure users find it easier to give feedback on all the insights they see on the dashboard and to use this feedback to circle it back to the models for further hypertuning thus improving accuracy.
To Begin with...
C
O
L
L
E
C
T
I
N
G
F
E
D
B
N
A
L
Y
S
I
N
G
F
E
E
D
B
C
K
A
C
T
O
N
F
E
E
D
B
C
K
FEEDBACK
LOOP
O
L
L
O
W
U
P
Scenario
The Cropin PR team needs to validate the accuracy of the data displayed on the dashboard against the actual field conditions. They request a field agent to collect data directly from the field and report back.
“
“
How are we collecting feedback now?
Through Elaborated Forms and Excel Sheets 😣
Some problems in this approach
🕐
Compulsory One-Time Data Entry
Field agents were asked to collect and input all data at once using a single form, leading to overwhelming tasks and potential errors if rushed.
⭐️
Subjective Rating Issues
Field agents were asked to rank items on a scale of 1 to 10. However, the interpretation of ratings varied, making it difficult to understand what a rating of 5 meant to us and them.
😮💨
User Fatigue and Incomplete Submissions
Users were forced to write detailed explanations of what went wrong, leading to user fatigue and resulting in incomplete data submissions as they rushed to finish.
📊
Limited Customization for Complex Data Needs
These Forms offered limited customization options for complex data collection, restricting the ability to capture all necessary data accurately.
Now lets get into the deetss...
Data Models on which feedback has to be collected
Predictive
DEWS (Advisory)
Deductive
Indices Model
Deductive
Weather & Climate Model
Predictive
Yield Model
Health Indicators
Crop Greenness
Health Indicators
Nitrogen Uptake
Health Indicators
Water Stress
Raw Indices
NDVI
Raw Indices
NDRE
Raw Indices
LSWI
What all information do we need to capture and collect from the user to make sense of the feedback and improve it?
Attributes we need to capture through the system
What is the Location?
GPS
What is the Zone?
GPS
Which Project?
Mixpanel
Who is the Farmer?
Mixpanel
What Crop is growing on the plot?
Mixpanel
At what Time the feedback was collected?
Mixpanel
What Plot is it field agent is collecting feedback on?
Mixpanel
What Specific Index, the feedback is collected?
Mixpanel
Attributes we need to collect from the user
Images captured to validate the on field conditions
Data Accuracy Rating
Weather Conditions because it could cause the discrepencies.
Data Update Issues
Any anomalies that were noticed
Plant Health Issues, Soil Health, etc
Somethings to keep in mind
🔄
Ensure Consistency
Maintain a consistent design across the feature, even if the model data outputs vary significantly from one another.
🛜
Provide Offline Support
Implement features that allow users to give feedback without needing a constant internet connection.
📙
Simplify Feedback Submission:
Make it easier for users to provide feedback by offering predefined answers and other helpful options.
📊
Facilitate Data Comprehension
Design the feedback system to enhance our ability to understand and analyze the collected data effectively.
Entry Points for Collecting Feedback
Crop Health - Collecting Feedback
Collecting Feedback on Overall Field Health
Collecting Feedback on Specific Points on the Field
Crop Stage Progression/Harvest - Collecting Feedback
Disease Warnings/Weather - Collecting Feedback
Success Metrics
User Experience Metrics
🔖
Task Completion rate
To Increase the efficiency in the feedback workflow.
⏳
Time on Task
To decreased the time spent navigating and giving the feedback on data models.
🙅🏻♂️
Error Reduction
To minimize user input error and confusion, leading to higher data accuracy of data models
Business Metrics
✅
To give accurate insights to users and ensure the feedback is used to improve the model insights on a continuous basis
📈
To increase user engagement
Learnings
A deep understanding of user needs and pain points guided the success of this design
Analysis of user behaviour has provided valuable insights which we can later build on. Collaborating with Product Managers, I’ve delved into understanding about the usage of the app
It is crucial to understand the feasibility of the design concepts and to align with the developers to implement the complex workflows.
Managing this project presented challenges, particularly in identifying edge cases and handling customer feedbacks. this experience has honed my skills as a designer and improved my collaboration with both the technical and product teams
Planning of Next Phases
The Report to be followed back to the agro-manager on the web application after looking at the success of this feature
To get the feedback on Product and App Features from the users.
