03 · DATA TOOLING
3–4 years of data, one screen, zero budget
CONTEXT
Three to four years of influencer group-buying campaign data was stored across individual Excel files. Retrieving performance data for a specific campaign or tracking cumulative creator contribution required manually opening and consolidating multiple files each time — an inefficient structure that created unnecessary friction in decision-making.
PROBLEM
The data existed but was not usable in its current state. Repeated analyses — campaign performance comparison, creator contribution tracking — were being done manually every time. A solution was needed that required no new tool procurement or development resources.
MY ROLE
Contribution: 100%. Problem identification, tool selection, data cleaning, dashboard design, and build — all completed independently. A self-initiated project, started during overtime without any external directive.
APPROACH
An underutilized dashboard feature within an existing internal tool was identified. It was implementable without development resources, and work began immediately. The project was executed in three stages.
Data Collection and Integration
Distributed Excel files were consolidated into a single data source. Inconsistent formats and creator name variations across files were standardized under a unified schema.
Data Cleaning
Creator name discrepancies, missing values, and duplicate entries were resolved manually. Input format standards were established to ensure future data would accumulate consistently.
Visualization
A dashboard was built around the most frequently needed metrics: GMV by campaign, cumulative contribution by creator, and performance by category.
RESULT
Manual retrieval and consolidation for campaign performance and creator contribution tracking was replaced by a single dashboard view. The tool saw active internal use until AI agent-based workflow automation became more widely adopted.
LESSON
The value of data is maximized through cleaning and structuring, not collection alone. If the format in which data is recorded is not standardized, even the best analysis tools will fail to deliver usable output. This project confirmed that standardizing data entry is a prerequisite for any analytics infrastructure.
No formal onboarding was provided to the team after the dashboard was built. Knowing a tool exists and actually using it for decision-making are different things — closing that gap should have been part of the rollout.