DEV Community: DanielNnadi

# dbt + Snowflake: From 45 Minutes to 3 Minutes Refresh

DanielNnadi — Sat, 13 Jun 2026 11:41:37 +0000

Excel reporting took 45 minutes to refresh and was error-prone. No version control. No data quality tests.

I migrated to dbt + Snowflake.

What I Built

15 Excel reports → dbt models
Star schema design
Incremental models for 2.3M+ rows
40+ automated data quality tests
CI/CD with GitHub Actions

Results

Metric	Before	After
Refresh time	45 minutes	3 minutes
Data quality	Manual	40+ tests
Anomalies caught	0%	12%

🔗 [View my portfolio]
https://danielnnadi.github.io/DanielNnadi/

Lean Six Sigma: How I Cut Manual Work from 40 to 14 Hours/Week

DanielNnadi — Sat, 13 Jun 2026 11:32:11 +0000

Weekly manual effort was 40 hours. Supplier on-time delivery was only 87%.

I applied Lean Six Sigma methodology to fix both.

What I Did

Mapped current processes (value stream mapping)
Identified 7 types of waste
Eliminated redundant approvals
Automated manual data entry
Built Power BI supplier dashboard

Results

Metric	Before	After
Weekly manual effort	40 hours	14 hours
Supplier on-time	87%	98%

🔗 [View my portfolio]
https://danielnnadi.github.io/DanielNnadi/

# How I Got 95% User Adoption on a New Inventory System

DanielNnadi — Sat, 13 Jun 2026 11:26:51 +0000

Vendors wanted to build 50+ features. After talking to 50+ staff members, I found they just wanted accurate data.

I cut the feature list by 30% before writing any code.

What I Did

User research with 50+ staff members
Cut vendor feature list by 30%
Defined API contracts before development
Built SQL architecture and Power BI dashboards

The Result

95% user adoption within 30 days
82% faster queries (45s → 8s)
On-time delivery

Key Takeaway

Data consistency matters more to users than feature quantity. Research before building.

🔗 [View my portfolio]
https://danielnnadi.github.io/DanielNnadi/

How I Migrated 10,000+ Records with Zero Downtime

DanielNnadi — Sat, 13 Jun 2026 11:21:24 +0000

At the Federal Ministry of Works, I led a data migration across 4 government directorates. The requirement: zero data loss, zero downtime, zero audit findings.

The Challenge

10,000+ records
4 different directorates
No system downtime allowed
Government audit requirements

The Solution

Saturday Cutover Plan:

Time	Action
6:00 AM	Start migration
7:00 AM	Backup source data
11:00 AM	Load to new system
2:00 PM	Switch over

Rollback ready at every step.

Results

Zero data loss
Zero downtime
Zero audit findings
Audit prep: 3 weeks → 3 days

🔗 [View my portfolio]
https://danielnnadi.github.io/DanielNnadi/

# How I Cut SQL Query Time from 45 Seconds to 8 Seconds

DanielNnadi — Sat, 13 Jun 2026 11:18:13 +0000

When I joined the data team at an automotive parts company, the inventory query everyone relied on took 45 seconds to run. Real-time reporting was impossible.

This is how I brought that query down to 8 seconds.

The Problem

The database had an inventory table with 2.3M+ rows. The query was doing full table scans on every execution.

The Fix

1. Read the execution plan - Found correlated subqueries running 500+ times

2. Replaced subquery with CTE:
Project Link


sql
- Before: 45 seconds
SELECT * FROM inventory i
WHERE (SELECT MAX(date) FROM orders o 
       WHERE o.sku_id = i.sku_id) IS NOT NULL
 - After: 8 seconds
WITH latest_orders AS (
    SELECT sku_id, MAX(date) as last_order
    FROM orders GROUP BY sku_id
)
SELECT i.* FROM inventory i
JOIN latest_orders lo ON i.sku_id = lo.sku_id

How I Built a Python Demand Forecasting Model That Saved $18,000

DanielNnadi — Sat, 13 Jun 2026 10:58:46 +0000

Emergency procurement was happening 23 times per month at the automotive parts company where I work. The purchasing team was reactive rather than proactive, causing stockouts and premium shipping costs.

I built a demand forecasting model that reduced emergency procurement by 65% and saved $18,000 annually.

The Problem

The team was restocking after stock hit zero rather than before. By the time they noticed, supplier lead times meant there was already a gap. Emergency orders cost more and disrupted operations.

The Solution

Data Collection

I pulled 13 months of stock movement data from our SQL inventory database — roughly 195,000 rows across 512 SKUs.

Feature Engineering

I engineered features including:

Rolling average consumption (7-day and 30-day)
Days of stock remaining
Stock buffer ratio
Seasonal indicators (month, day of week)

Model Selection

I tested three approaches:

Simple threshold rule (too many false positives)
Gradient Boosting (good but lower precision)
Random Forest (best balance)

Random Forest gave 78% precision and 76% recall.

Deployment

The model runs daily at 7 AM and outputs a prioritized reorder list directly to the operations team as an Excel report.

Results

Metric	Before	After
Emergency orders/month	23	8
Reduction	-	65%
Annual savings	-	$18,000

Key Takeaway

The model's accuracy mattered less than the cross-functional review process. Consistent weekly meetings between procurement and data teams drove the real behavior change.
project link

Connect With Me

📊 Portfolio:
(https://danielnnadi.github.io/DanielNnadi/)

💼 LinkedIn: linkedin.com/in/danielnnadi

🐙 GitHub: github.com/danielnnadi