Zacharias CuraData Analyst · Data Engineering

I build and validate e‑commerce data products.

At Great Deals E-Commerce, I work with marketplace data from Shopee, Lazada, and TikTok. I build Snowflake/dbt pipelines, reconcile Seller Center exports with Python, investigate QuickSight reporting discrepancies by tracing them back to SQL, and build Streamlit pages for sales-conversion analytics.

View selected work View experience Download CV

Based in the Philippines · Growing from data analysis into analytics and data engineering

SnowflakedbtSQLPython / pandasAirflowAWS QuickSightStreamlitPowerShellGit / GitHub

Selected work

Real problems, decisions, and validation.

These projects are written for both technical and non-technical readers: what was wrong, what I changed, how I checked it, and what improved.

About the examplesCompany and client details are generalized. Any sample rows or values shown are synthetic and labeled; the project results and counts are from my work.
Synthetic sample
Order / SKUSeller CenterSnowflakeClassification
SP-1042 / A12PaidPaidMatch
SP-1043 / B0710:1410:17Timing
SP-1044 / C31Complete—Needs review
SP-1045 / D09PresentGap dateSource gap

Shopee · Automation · Data quality

Shopee Seller Center Reconciliation

Built recon_engine.py to compare Shopee Seller Center exports with API-sourced Snowflake data, moving repeated row-level reconciliation out of Claude/ChatGPT and into a deterministic Python/pandas workflow.

Result: 107,160 rows compared without using LLM tokens for the core comparison, with exceptions classified for focused investigation.

PythonpandasSnowflakeReconciliation
Explore project →
Sanitized architecture
Bronzesource-aligned modelsdbt
Silvercleaned business logicincremental
Goldunified funnel grainBI-ready
Airflow orchestration → QuickSight + Streamlit

dbt · Analytics engineering

Sales Conversion Data Pipeline

Converted sales-conversion SQL into a maintained Snowflake/dbt pipeline with a unified Gold reporting grain, incremental processing, and scheduled orchestration.

Result: 13 dbt models across Bronze, Silver, and Gold, feeding QuickSight and Streamlit analytics.

SnowflakedbtAirflowStreamlit
Explore project →
Synthetic sample
Before
158
account target repeated across employee rows
→
After grain fix
100
target rolls up at the intended grain

QuickSight · Root-cause analysis

Correcting an Overstated KPI

Traced a QuickSight KPI that was 58% too high to SQL that repeated the full account target across employee rows, then rebuilt the reporting logic so targets rolled up correctly.

Result: Corrected the 58% overstatement by aligning account targets with the employee reporting grain before aggregation.

SQLQuickSightSnowflakeValidation
Explore project →

Lazada · Shopee · TikTok

Platform-fee pipeline

Built dbt Silver and Gold models in Snowflake from Airflow-ingested Bronze platform-fee data, standardizing and deduplicating marketplace records into an effective-dated SCD Type 2 dimension with SHA2-256 keys, data-quality flags, and dbt tests.

Result: three marketplace fee sources now follow one modeled structure for downstream analytics.

Explore project →

Lazada · SQL

77-column order report

Mapped the 77-column Seller Center report scope into one optimized Snowflake query joining six tables across two databases, using pre-aggregation CTEs and filter pushdown to control fan-out and compute cost.

Validation: 39 columns reached exact match against Seller Center; remaining field differences were documented and tracked.

Explore project →

Automation · Internal tooling

EMS Mailbox Provisioning Automation

Replaced a manual process where each EMS mailbox was created one at a time and its credentials were then saved to Passbolt manually. The PowerShell workflow now handles mailbox creation and credential storage end to end.

Result: a 30-account batch went from approximately 150 minutes of manual work to about 2 minutes with automation.

Explore project →

Experience

Working across marketplace data, BI, automation, and Snowflake.

Apr 2026 — Present

Junior Data Analyst

Great Deals E-Commerce Corporation

  • Built Snowflake/dbt Silver and Gold models from Airflow-ingested Bronze platform-fee data, standardizing Lazada, Shopee, and TikTok records into an effective-dated SCD Type 2 dimension with SHA2-256 keys, data-quality flags, and dbt tests.
  • Built recon_engine.py in Python/pandas to compare Shopee Seller Center exports with API-sourced Snowflake data; compared 107,160 rows and classified timing, missing-record, and source-coverage exceptions for review.
  • Built a sales-conversion Gold model for QuickSight and Streamlit, then converted the workflow into 13 dbt models with incremental processing and Airflow orchestration.
  • Reverse-engineered a 77-column Lazada report into an optimized Snowflake query across six tables and two databases.
  • Corrected a QuickSight KPI overstated by 58% when account targets repeated across employee rows; rebuilt the Snowflake reporting view and allocated monthly account targets across assigned employees for accurate roll-ups.
  • Automated EMS mailbox provisioning and Passbolt credential storage using PowerShell, REST APIs, Passbolt CLI, and mTLS, reducing a 30-account batch from approximately 150 minutes of manual work to about 2 minutes.
Snowflake governance: I currently handle operational access work under SECURITYADMIN, including provisioning/deprovisioning and network rules. A broader review of unmanaged users, roles, grants, and role hierarchy is an ongoing governance initiative rather than completed cleanup.
Jul 2024 — Apr 2026

Surveillance Operator

Solaire Resort North

  • Managed 30–80 inbound calls per shift while monitoring gaming and cash-handling operations across multiple systems and producing structured compliance and incident records.

About me

I like finding where the data stops matching the business.

I’m a Computer Engineering graduate working as a Junior Data Analyst in e-commerce. Most of my day-to-day work is hands-on: comparing marketplace exports with warehouse data, turning SQL into dbt models, debugging BI numbers, and automating repeatable operational tasks.

I’m moving deeper into analytics and data engineering, especially warehouse modeling, data quality, orchestration, governance, and reliable data products.

Skills

Tools I use in the work above.

Data engineering & ETL

dbt, Airflow, Python (pandas), SQL, medallion architecture, SCD Type 2, dimensional modeling, incremental processing, data-quality testing

Warehousing & governance

Snowflake, SECURITYADMIN workflows, roles and grants, network rules, external stages/tables, PostgreSQL

BI & data applications

AWS QuickSight, Streamlit, Power BI, Tableau, SPICE, calculated fields, KPI validation

Automation & tooling

PowerShell, REST APIs, Passbolt CLI, mTLS, Git, GitHub, VS Code, WSL

Credentials

Education and certifications.

Computer Engineering foundations supported by focused training in analytics and SQL.

Education · 2021–2024

Bachelor of Science in Computer Engineering

STI College Ortigas – Cainta

Certification

Data Analytics Essentials

Cisco Networking Academy

Certification

The Complete SQL Bootcamp

Udemy · Pierian Training

Contact

Interested in the work behind the numbers?

I’m open to conversations around data analyst, analytics engineering, and data engineering opportunities.

curazacharias@gmail.comLinkedInGitHubDownload CV