Gema Amran Nathansyach
Data Analyst
I help businesses make sense of their data, especially in e-commerce: sales trends, customer retention, and delivery performance. I work with SQL, Python (pandas, Matplotlib), and Excel.
Work
Below are some fruits of my thought
SQL + Python
Olist E-Commerce Sales Analysis
Background
Olist is a Brazilian online marketplace that connects small sellers with buyers all over the country. I looked at how it sells and what its customers experience.
Purpose
Find what sells best, see how well deliveries go, and find ways to bring customers back.
Business questions
- Is the business growing?
- What sells the most?
- Do late deliveries hurt reviews?
- Do customers come back?
- Do a few sellers bring in most of the money?
Data
- Orders
- Order items
- Payments
- Customer reviews
- Products
- Sellers
Stack
Sales overview
Jan 2017 — Aug 2018Monthly revenue
BRLReview vs. delivery
avg. score / 5Delivery time by state
avg. days- São Paulo (SP)8.8
- Minas Gerais (MG)12.0
- Rio de Janeiro (RJ)15.3
- Bahia (BA)19.3
- Pará (PA)23.8
- Amazonas (AM)26.4
- Roraima (RR)29.4
Top categories
% of revenue- Health & beauty9.3%
- Watches & gifts8.8%
- Bed, bath & table7.7%
- Sports & leisure7.2%
- Computers6.7%
Lowest-rated categories
avg. score / 5- Office furniture3.52
- Bed, bath & table3.92
- Furniture & decor3.95
- Computers3.99
- Telephony3.99
Revenue by seller group
% of revenuePayment mix
% of value- Credit card78.5%
- Boleto18.0%
- Voucher2.2%
- Debit card1.3%
Card buyers split payments into 3.5 installments on average.
Repeat customers
% of customers- 1 order97.0%
- 2 orders2.8%
- 3+ orders0.2%
Only 2,801 of 93,350 customers ordered more than once.
Data checks
found before analysis- Orders not delivered2,963
- Repeat buyers hidden by customer ID3,345
- Orders without any item775
- Products without a category610
- Orders with 2+ reviews547
- Payment ≠ items + freight249
- Delivered, but no delivery date8
License
Experience
I've worked on data analysis, research, web development, and my own small business. In all of them, I use data to make better decisions. Here's a quick look.
Nov 2025 — Now
- Olist sales analysis: I analyzed 100K orders from Olist, a Brazilian marketplace, to find where the business is doing well and where it's losing money.
- SQL and Python: I combined 8 tables with SQL (JOINs, window functions) and Python (pandas, Matplotlib), and wrote down every data cleaning decision.
- Key findings: deliveries just 1–3 days late lowered reviews by a full star. 97% of customers only bought once, and the top 10% of sellers made 67% of the revenue.
- Recommendations: I turned the findings into 5 next steps, ordered by priority, covering delivery, customer retention, and sellers.
Nov 2025
Thesis
Mar 2025 — Feb 2026
Jun 2024 — Feb 2026
