Hot Topic: Data Dictionary

DigiCow Farmer Training Adoption Challenge

Helping Kenya

€8 250 EUR

Under code review

Skills you will learn

Data analysis

Classification

895 joined

388 active

Info Data Chat Leaderboard

Start

Jan 28, 26

Mar 01, 26

Reveal

Mar 02, 26

Usurper

Data Dictionary

Data · 10 Feb 2026, 07:51 · 5

I am struggling to understand some explanations for features in the data dictionary, specifically: has_topic_traine_on - its description does not make sense, perhaps some words are missing, and topics: are these topics the farmer has already been trained on, or the potential ones they could be trained on?

Would appreciate some help

Discussion 5 answers

Juliuss

Freelance

yes this is a valid issue raised..@AJoel could you also shed more light on feature group_name? I see it in the datasets but no mention of it in the data dictionary.

11 Feb 2026, 21:52

Upvotes 0

AJoel

Zindi

Some farmers belong to cooperative and are organised into group. So that is what the group_name refers to.

replied to Juliuss11 Feb 2026, 23:07

Upvotes 1

Juliuss

Freelance

Perfect👌.

Thanks so much for the feedback @Ajoel

replied to AJoel12 Feb 2026, 05:05

Upvotes 0

Juliuss

Freelance

@Ajoel, these are my observations and assumptions based on your response

Observations;

Most farmers do not belong to a cooperative. 89.2% in train and about 96.9% in test
group_name has no nulls-each farmer has a group still despite them not belonging to a cooperative

Assumptions:

Either everyone is assigned some default group
Or “non-cooperative” farmers are still grouped administratively

Incase I am wrong with the assumptions you can correct

replied to Juliuss12 Feb 2026, 05:34

Upvotes 0

AJoel

Zindi

Hello: has_topic_trained_on is a boolean flag. It is 1 if the specific topic that the farmer was trained on is recorded. It should be has_topic_trained_on instead of "has_topic_traine_on"

11 Feb 2026, 23:06

Upvotes 1

Join the largest network for
data scientists and AI builders

About FAQs

Status